Big Data Analytics for Satellite Image

Processing and Remote Sensing

P. Swarnalatha
VIT University, India

Prabu Sevugan
VIT University, India

A volume in the Advances in Computer and Electrical Engineering (ACEE)

Book Series Book Series
Published in the United States of America by IGI Global (an imprint of IGI
Global) 701 E. Chocolate Avenue Hershey PA 17033 Tel: 717-533-8845 Fax:
717-533-8661 E-mail: [email protected] Web site: http://www.igi-
global.com/reference

Copyright © 2018 by IGI Global. All rights reserved. No part of this publication
may be reproduced, stored or distributed in any form or by any means, electronic
or mechanical, including photocopying, without written permission from the
publisher.

Product or company names used in this set are for identification purposes only.
Inclusion of the names of the products or companies does not indicate a claim of
ownership by IGI Global of the trademark or registered trademark.

Library of Congress Cataloging-in-Publication Data

Names: Swarnalatha, P. (Purushotham), 1977-editor. | Sevugan, Prabu, 1981-

editor.
Title: Big data analytics for satellite image processing and remote sensing /
P. Swarnalatha and Prabu Sevugan, editors.
Description: Hershey, PA : Engineering Science Reference, [2018] | Includes
bibliographical references.
Identifiers: LCCN 2017019875| ISBN 9781522536437 (hardcover) | ISBN
9781522536444 (ebook)
Subjects: LCSH: Geospatial data. | Big data. | Artificial satellites in
remote sensing. | Image processing--Digital techniques. | Artificial
satellites--Optical observations--Data processing.
Classification: LCC GA102.4.E4 B54 2018 | DDC 621.36/7028557--dc23 LC
record available at https://lccn.loc.gov/2017019875

This book is published under the IGI Global book series Advances in Computer
and Electrical Engineering (ACEE) (ISSN: 2327-039X eISSN: 2327-0403)
British Cataloguing in Publication Data

A Cataloguing in Publication record for this book is available from the British
Library.

All work contributed to this book is new, previously-unpublished material. The

views expressed in this book are those of the authors, but not necessarily of the
publisher.
Advances in Computer and
Electrical Engineering (ACEE) Book
Series
Srikanta Patnaik (SOA University, India)

ISSN: 2327-039X

Mission

The fields of computer engineering and electrical engineering encompass a

broad range of interdisciplinary topics allowing for expansive research
developments across multiple fields. Research in these areas continues to
develop and become increasingly important as computer and electrical systems
have become an integral part of everyday life.

The Advances in Computer and Electrical Engineering (ACEE) Book Series

aims to publish research on diverse topics pertaining to computer engineering
and electrical engineering. ACEE encourages scholarly discourse on the latest
applications, tools, and methodologies being implemented in the field for the
design and development of computer and electrical systems.

Coverage IGI Global is currently

accepting manuscripts
for publications within
Analog Electronics this series. To submit a
Qualitative Methods
Power Electronics
proposal for a volume
Programming in this series please
Digital Electronics contact our
Electrical Power Acquisition Editors at
Conversion
Algorithms Acquisitions@igi-
Computer global.com or visit
 Architecture http://www.igi-
Computer Hardware
VLSI Design
global.com/publish .

The Advances in Computer and Electrical Engineering (ACEE) Book

Series(ISSN 2327-039X) is published by IGI Global, 701 E. Chocolate Avenue,
Hershey, PA 17033-1240, USA, www.igi-global.com. This series is composed
of titlesavailable for purchase individually; each title is edited to be contextually
exclusive from any other title within the series. For pricing and ordering
information please visit http://www.igi-global.com/bookseries/advances-
computer-electrical-engineering/73675. Postmaster: send all address changes to
above address. Copyright © 2018 IGI Global. All rights, including translation in
other languages reserved by the pulisher. No part of this series may be
reproduced or used in any form or by any means - graphics, electronic, or
mechanical, including photocopying, recoreding, taping, or information and
retrieval systems - without written permission from the publisher, except for non
commercial, educational use, including classroom teaching purposes. The views
expressed in this series are those of the authors, but not necessarily of IGI Global

Titles in this Series

EHT Transmission Performance Evaluation Emerging Research and
Opportunities
K. Srinivas (Transmission Corporation of Andhra Pradesh Limited, India) and
R.V.S. Satyanarayana (Sri Venkateswara University College of Engineering,
India)
Engineering Science Reference • copyright 2018 • 160pp • H/C (ISBN:
9781522549413) • US $145.00 (our price)

Fuzzy Logic Dynamics and Machine Prediction for Failure Analysis

Tawanda Mushiri (University of Johannesburg, South Africa) and Charles
Mbowhwa (University of Johannesburg, South Africa)
Engineering Science Reference • copyright 2018 • 301pp • H/C (ISBN:
9781522532446) • US $225.00 (our price)

Creativity in Load-Balance Schemes for Multi/Many-Core Heterogeneous

Graph Computing Emerging Research and Opportunities
Alberto Garcia-Robledo (Center for Research and Advanced Studies of the
National Polytechnic Institute (Cinvestav-Tamaulipas), Mexico) Arturo Diaz-
Perez (Center for Research and Advanced Studies of the National Polytechnic
Institute (Cinvestav-Tamaulipas), Mexico) and Guillermo Morales-Luna (Center
for Research and Advanced Studies of the National Polytechnic Institute
(Cinvestav-IPN), Mexico)
Engineering Science Reference • copyright 2018 • 217pp • H/C (ISBN:
9781522537991) • US $155.00 (our price)

Free and Open Source Software in Modern Data Science and Business
Intelligence Emerging Research and Opportunities
K.G. Srinivasa (CBP Government Engineering College, India) Ganesh Chandra
Deka (M. S. Ramaiah Institute of Technology, India) and Krishnaraj P.M. (M. S.
Ramaiah Institute of Technology, India)
Engineering Science Reference • copyright 2018 • 189pp • H/C (ISBN:
9781522537076) • US $190.00 (our price)

Design Parameters of Electrical Network Grounding Systems

Osama El-Sayed Gouda (Cairo University, Egypt)
Engineering Science Reference • copyright 2018 • 316pp • H/C (ISBN:
9781522538530) • US $235.00 (our price)

Design and Use of Virtualization Technology in Cloud Computing

Prashanta Kumar Das (Government Industrial Training Institute Dhansiri, India)
and Ganesh Chandra Deka (Government of India, India)
Engineering Science Reference • copyright 2018 • 315pp • H/C (ISBN:
9781522527855) • US $235.00 (our price)

Smart Grid Test Bed Using OPNET and Power Line Communication
Jun-Ho Huh (Catholic University of Pusan, South Korea)
Engineering Science Reference • copyright 2018 • 425pp • H/C (ISBN:
9781522527763) • US $225.00 (our price)

Transport of Information-Carriers in Semiconductors and Nanodevices

Muhammad El-Saba (Ain-Shams University, Egypt)
Engineering Science Reference • copyright 2017 • 677pp • H/C (ISBN:
9781522523123) • US $225.00 (our price)

Accelerating the Discovery of New Dielectric Properties in Polymer Insulation

Boxue Du (Tianjin University, China)
Engineering Science Reference • copyright 2017 • 388pp • H/C (ISBN:
9781522523093) • US $210.00 (our price)

Handbook of Research on Nanoelectronic Sensor Modeling and Applications

Mohammad Taghi Ahmadi (Urmia University, Iran) Razali Ismail (Universiti
Teknologi Malaysia, Malaysia) and Sohail Anwar (Penn State University, USA)
Engineering Science Reference • copyright 2017 • 579pp • H/C (ISBN:
9781522507369) • US $245.00 (our price)

Field-Programmable Gate Array (FPGA) Technologies for High Performance

Instrumentation
Julio Daniel Dondo Gazzano (University of Castilla-La Mancha, Spain) Maria
Liz Crespo (International Centre for Theoretical Physics, Italy) Andres Cicuttin
(International Centre for Theoretical Physics, Italy) and Fernando Rincon Calle
(University of Castilla-La Mancha, Spain)
Engineering Science Reference • copyright 2016 • 306pp • H/C (ISBN:
9781522502999) • US $185.00 (our price)

Design and Modeling of Low Power VLSI Systems

Manoj Sharma (BVC, India) Ruchi Gautam (MyResearch Labs, Gr Noida, India)
and Mohammad Ayoub Khan (Sharda University, India)
Engineering Science Reference • copyright 2016 • 386pp • H/C (ISBN:
9781522501909) • US $205.00 (our price)

Reliability in Power Electronics and Electrical Machines Industrial Applications

and Performance Models
Shahriyar Kaboli (Sharif University of Technology, Iran) and Hashem Oraee
(Sharif University of Technology, Iran)
Engineering Science Reference • copyright 2016 • 481pp • H/C (ISBN:
9781466694293) • US $255.00 (our price)

Handbook of Research on Emerging Technologies for Electrical Power

Planning, Analysis, and Optimization
Smita Shandilya (Sagar Institute of Research Technology & Science, India)
Shishir Shandilya (Bansal Institute of Research & Technology, India) Tripta
Thakur (Maulana Azad National Institute of Technology, India) and Atulya K.
Nagar (Liverpool Hope University, UK)
Engineering Science Reference • copyright 2016 • 410pp • H/C (ISBN:
9781466699113) • US $310.00 (our price)
Sustaining Power Resources through Energy Optimization and Engineering
Pandian Vasant (Universiti Teknologi PETRONAS, Malaysia) and Nikolai
Voropai (Energy Systems Institute SB RAS, Russia)
Engineering Science Reference • copyright 2016 • 494pp • H/C (ISBN:
9781466697553) • US $215.00 (our price)

Environmental Impacts on Underground Power Distribution

Osama El-Sayed Gouda (Cairo University, Egypt)
Engineering Science Reference • copyright 2016 • 405pp • H/C (ISBN:
9781466665095) • US $225.00 (our price)

Handbook of Research on Advanced Computational Techniques for Simulation-

Based Engineering
Pijush Samui (Centre for Disaster Mitigation and Management, VIT University,
India)
Engineering Science Reference • copyright 2016 • 616pp • H/C (ISBN:
9781466694798) • US $310.00 (our price)

Operation, Construction, and Functionality of Direct Current Machines

Muhammad Amin (COMSATS Institute of Information Technology, Wah Cantt,
Pakistan) and Mubashir Husain Rehmani (COMSATS Institute of Information
Technology, Wah Cantt, Pakistan)
Engineering Science Reference • copyright 2015 • 404pp • H/C (ISBN:
9781466684416) • US $200.00 (our price)

Performance Optimization Techniques in Analog, Mixed-Signal, and Radio-

Frequency Circuit Design
Mourad Fakhfakh (University of Sfax, Tunisia) Esteban Tlelo-Cuautle (INAOE,
Mexico) and Maria Helena Fino (New University of Lisbon, Portugal)
Engineering Science Reference • copyright 2015 • 464pp • H/C (ISBN:
9781466666276) • US $235.00 (our price)

Agile and Lean Service-Oriented Development Foundations, Theory, and

Practice
Xiaofeng Wang (Free University of Bozen/Bolzano, Italy) Nour Ali (Lero-The
Irish Software Engineering Research Centre, University of Limerick, Ireland)
Isidro Ramos (Valencia University of Technology, Spain) and Richard Vidgen
(Hull University Business School, UK)
Information Science Reference • copyright 2013 • 312pp • H/C (ISBN:
9781466625037) • US $195.00 (our price)

Software Testing in the Cloud Perspectives on an Emerging Discipline

Scott Tilley (Florida Institute of Technology, USA) and Tauhida Parveen
(Independent Consultant, USA)
Information Science Reference • copyright 2013 • 479pp • H/C (ISBN:
9781466625365) • US $195.00 (our price)

Electromagnetic Transients in Transformer and Rotating Machine Windings

Charles Q. Su (Charling Technology, Australia)
Engineering Science Reference • copyright 2013 • 586pp • H/C (ISBN:
9781466619210) • US $195.00 (our price)

Handbook of Research on Computational Science and Engineering Theory and

Practice
J. Leng (Visual Conclusions, UK) and Wes Sharrock (University of Manchester,
UK)
Engineering Science Reference • copyright 2012 • 987pp • H/C (ISBN:
9781613501160) • US $495.00 (our price)

Dependability and Computer Engineering Concepts for Software-Intensive

Systems
Luigia Petre (Åbo Akademi University, Finland) Kaisa Sere (Åbo Akademi
University, Finland) and Elena Troubitsyna (Åbo Akademi University, Finland)
Engineering Science Reference • copyright 2012 • 515pp • H/C (ISBN:
9781609607470) • US $195.00 (our price)

Design and Test Technology for Dependable Systems-on-Chip

Raimund Ubar (Tallinn University of Technology, Estonia) Jaan Raik (Tallinn
University of Technology, Estonia) and Heinrich Theodor Vierhaus
(Brandenburg University of Technology Cottbus, Germany)
Information Science Reference • copyright 2011 • 578pp • H/C (ISBN:
9781609602123) • US $180.00 (our price)

Modern Software Engineering Concepts and Practices Advanced Approaches

Ali H. Dogru (Middle East Technical University, Turkey) and Veli Biçer (FZI
Research Center for Information Technology, Germany)
Information Science Reference • copyright 2011 • 506pp • H/C (ISBN:
9781609602154) • US $180.00 (our price)
Kansei Engineering and Soft Computing Theory and Practice
Ying Dai (Iwate Pref. University, Japan) Basabi Chakraborty (Iwate Prefectural
University, Japan) and Minghui Shi (Xiamen University, China)
Engineering Science Reference • copyright 2011 • 436pp • H/C (ISBN:
9781616927974) • US $180.00 (our price)

Model Driven Architecture for Reverse Engineering Technologies Strategic

Directions and System Evolution
Liliana Favre (Universidad Nacional de Centro de la Proviencia de Buenos
Aires, Argentina)
Engineering Science Reference • copyright 2010 • 461pp • H/C (ISBN:
9781615206490) • US $180.00 (our price)

Strategies for Managing Computer Software Upgrades

Neil G. Shaw (University of Texas-Arlington, USA)
IGI Publishing • copyright 2001 • 260pp • H/C (ISBN: 9781930708044) • US
$69.95 (our price)

701 E. Chocolate Ave., Hershey, PA 17033

Order online at http://www.igi-global.com or call 717-533-8845 x100
To place a standing order for titles released in this series, contact:
[email protected]
Mon-Fri 8:30 - 5:00 pm (est) or fax 24 hours a day 717-533-8661
Foreword
S. Arunkumar Sangaiah
VIT University, India

When I was invited to write a foreword for the book Big Data Analytics for
Satellite Image Processing and Remote Sensing, I felt glad to note the varied
tools, challenges, methods in Bigdata for Satellite Image processing. This book
is a significant collection of 10 chapters covering image processing, satellite
image processing, bigdata and cloud based processing, as well as their
applications in emerged in the recent decades. This book provides an excellent
platform to review various areas of Satellite Image processing and affords for the
needs of both beginners to the field and seasoned researchers and practitioners.
The tremendous growth of Satellite Image Processing and Bigdata are
documented in this book, such as Bigdata, Satellite Image Processing, Remote
Sensing, Computational methods, 3D Asset development and product
development, Landslide susceptibility, Hierarchical clustering, Modified Support
Vector Machine, Bigdata as a Service and Cloud Based Workflow Scheduling
techniques which are focused in various applications.

To the best of my knowledge, this is the first attempt of its kind, providing a
coverage of the key subjects in the fields Bigdata, Satellite Image Processing and
Cloud Computing and applications. This book is an invaluable, topical, and
timely source of knowledge in the field, which serves nicely as a major text book
for several courses at both undergraduate, post graduate levels and scholars. It is
also a key reference for scientists, professionals, and academicians, who are
interested in new challenges, theories and practice of the specific areas
mentioned above.

I am happy to commend the editors and authors on their accomplishment, and to

inform the readers that they are looking at a major piece in the development of
computational intelligence on organizational decision making. I am familiar with
your research interests and expertise in the related research areas Big Data as a
Service, Big Data Industry Standards, Experiences with Big Data Project
Deployments. New Computational Models for image remote sensing and Big
Data which would make an excellent addition to this publication. This book is a
main step in this field’s maturation and will serve to challenge the academic,
main step in this field’s maturation and will serve to challenge the academic,
research and scientific community in various significant ways.
Preface
Gopinath Ganapathy
Bharathidasan University, India

V. Susheela Devi
Indian Institute of Science Bangalore, India

Ravee Sundararajan
London South Bank University (LSBU), UK

As we have entered an era of high resolution earth observation, the Remote

Sensing data are undergoing an explosive growth. The proliferation of data also
gives rise to the increasing complexity of remotely sensed data, like the diversity
and higher dimensionality characteristic of the data. This book aims to discuss
and address the difficulties and challenges faced in handling big data and
remotely sensed data applied in various applications. The editors have received
chapters that address different aspects of using big data upon satellite image
processing and related topics. Additionally, the book also explored the impact of
such methodologies on the applications in which this advanced technology is
being implemented.

This comprehensive and timely publication aims to be an essential reference

source. This book helps the researchers who are working on satellite image
processing with the available literature in the field of big data upon image
processing techniques on remote sensing while providing for further research
opportunities in this dynamic field. It is hoped that this text will provide the
resources necessary technology to the developers and managers to adopt and
implement these techniques, platforms and approaches in developing efficient
solutions.

NEED FOR A BOOK ON THE PROPOSED

TOPICS
Big data is an evolving term that describes any voluminous amount of
structured, semi-structured, and unstructured data that has the potential to be
structured, semi-structured, and unstructured data that has the potential to be
mined for meaningful information. The scope of image processing and
recognition has broadened due to the gap in scientific visualization. Thus, new
imaging techniques have developed, and it is imperative to study this
progression for optimal utilization. The data received from remote sensing
satellite in various forms has been used to study various applications in real
world by applying big data analytics.

Big Data Analytics for Satellite Image Processing and Remote Sensing is a
critical scholarly resource that examines the challenges and difficulties of
implementing big data in image processing for remote sensing and related areas.
Featuring coverage on a broad range of topics, such as distributed computing,
parallel processing, and spatial data, this book is geared towards scientists,
professionals, researchers, and academicians seeking current research on the use
of big data analytics in satellite image processing and remote sensing.

ORGANIZATION OF THE BOOK

The book is organized into 10 chapters. A brief description of each chapter is
given as follows:

1. New Computational Models for Image Remote Sensing and Big

Data: This chapter focuses on the development of new computational
models for remote sensing applications with big data handling method
using image data. Furthermore, this chapter presents overview of the
process of developing systems for remote sensing and monitoring. The
issues and challenges are presented to discuss various problems related to
the handling of image big data in wireless sensor network that has various
real-world applications. Moreover, the possible solutions and future
recommendations to address the challenges have been presented and also
this chapter includes discussion of emerging trends and conclusion.
2. Big Data Computation Model for Landslide Risk Analysis Using
Remote Sensing Data: Nowadays, effective and efficient strategies have to
acquire, manage and analyze data leads that to better decision making and
competitive advantage. The development of cloud computing and the big
data era, brings up challenges to traditional data mining algorithms. the
characteristics of complexity have to formulate an extreme challenge for
discovering useful knowledge from the big data. Spatial data is complex big
data. To handle complex spatial landslide data, this chapter is proposed
with Multi Ranking Decision Tree big data approach.. The Proposed
Classifier performance is validated with massive real time dataset. The
results indicate that our classifier exhibits both time efficiency and
scalability.
3. Modified Support Vector Machine Algorithm to Reduce
Misclassification and Optimizing Time Complexity: This chapter deals
with optimization of the time consumption for computation of classification
algorithm. All these results in reduction in miss-classification and faster
computation of class for data point. This caused due to some drawback of
each algorithm depending upon type of data. To reduce miss-classification
new methodology of “Support Vector Machine” is to be introduced, instead
of having hyperplane exactly in the middle position of hyper plane that has
to be changed per number of data points of class available near to the
hyperplane. Thereby, this chapter shows some multi core architecture
which has been used to compute more than one in dependent modules
simultaneously.
4. An Analysis of Usage-Induced Big Data: This chapter deals with the
“Analysis of Usage-induced Big Data”. The most effective method to store,
oversee, and make values from the administration situated enormous
information turn into a vital research issue. With the inexorably huge
measure of information, a solitary framework which gives normal
usefulness to overseeing and dissecting diverse sorts of administration
produced enormous information is critically required. This test is addressed
in this chapter which gives a review of administration produced huge
information and Big Data-as-a-Service that includes Big Data
Infrastructure-as-a-Service, Big Data Platform-as-a-Service, and Big Data
Analytics Software-as-a-Service, giving regular huge information related
administration (e.g., getting to benefit produced huge information and
information investigation results) to clients to improve effectiveness and
lessen cost.
5. Glorified Secure Search Over Encrypted Secured Cloud Database
Through Hierarchical Clustering Computation: Search on Cloud
Through Hierarchical Clustering: This chapter proposed an improvised
SE scheme to grasp the contest difficulties. In the development,
prototypical, hierarchical clustering technique is intended to leader
additional search semantics with a supplementary feature of making the
scheme to deal with the claim for reckless cipher text search in big scale
surroundings such situations where there is huge amount of data. This focus
on fetching of cloud data of a subcontracted encrypted information deprived
of loss of idea and if security and privacy by transmission attribute key to
the information with the next level the typical is improved with the
multilevel conviction privacy preserving scheme.
6. Research Analysis of Development Pipelines in Augmented and
Virtual Reality Technologies: The objective of this chapter is 3D asset
development and product development process and creation of real world
solutions using Augmented and Virtual Reality Technologies. The methods
involve creation of simulative software solutions that can be used in
assisting corporations with training activities as using Augmented Reality
(AR) and Virtual Reality (VR) training tools to cut costs. This chapter
applied AR and VR technologies for training purposes, a 60% cost
reduction can be observed. The application of AR and VR technologies can
help in using a smartphones, high performance computers, head mounted
displays (HMDs) other such technologies to provide solutions via
simulative environments. Thereby the chapter proved improvements by
implementing a good UX (user experience) in training reducing on-site
training risks and cut costs rapidly.
7. Iterative MapReduce: i-MapReduce on Medical Dataset Using
Hadoop: The goal of this chapter is to dissect and extricate the enhanced
performance of Data Analysis in distributed environment. Iterative
MapReduce (i-MapReduce) plays a major role in optimizing the analytics
performance by finding useful data patterns using reduced mapper scans.
The term 'Enormous Data' alludes to the monstrous volumes of both
organized and unstructured information which can’t be directly utilized
with conventional database administration frameworks. With the quick
increment of urbanization in India and its many-fold determinants, the
information will become tremendous and turns out to be Big Data which
couldn't be handled by traditional DBMS. In this chapter, Hadoop, an open
source structure that permits the disseminated handling for huge datasets on
group of PCs has been examined and thus finally produces better results
with the deployment of Iterative MapReduce.
8. Big Data for Satellite Image Processing: Analytics, Tools, Modeling,
and Challenges: This chapter presents an introduction to the basics in
bigdata including architecture, modeling and the tools used. Bigdata is a
term which is used for serving the high volume of data that can be used as
an alternative to RDBMS and the other analytical technologies such as
OLAP. In order to make it useful, the chapter has to deal it efficiently by
placing the bigdata that exceeds the processing and the overall capacity of
other traditional databases. This also discuss with the basic architecture,
tools, modeling and challenges to give efficient solution using bigdata for
satellite image processing.
9. Big Data Processing: Application of Parallel Processing Technique to
Big Data by Using MapReduce: This chapter deals with a description
about MapReduce which serves as a programming algorithm for distributed
computing in a parallel manner on huge chunks of data which can easily
execute on commodity servers. Thereby, reduce the costs of server
maintenance and removal of requirement of having dedicated servers
towards the running of processes. This chapter discuss about the various
approaches towards map reduce programming model and how to use it in
an efficient manner for scalable text based analysis in various domains like
machine learning, data analytics and data science. Hence it will deal with
various approaches of using map reduce in these fields and how to apply
various techniques of MapReduce in these fields effectively and fitting the
MapReduce programming model into any text mining application.
10. Image Processing in CES-4019 Distinction Between Fake Note and
a Real Note: In the chapter, the author is trying to distinguish between fake
note and a real note and would like to take it to a level where it can be used
everywhere. The data after the detection of the fakeness and the real note
can be stored in the database. As the data to store will be hug, to overcome
this problem, the chapter deals with Big Data. It will help us to store large
amount of data in no time. The difference between real note and fake note
is that real note has its thin strip to be more or less continuous while the
fake strip has fragmented thin lines in the strip. One could say that the fake
note has more than one line in the thin strip while the real note only has one
line. Therefore, the chapter deals with: just one line, it's real but if the users’
see more than one line it's fake. Here, the dataset used is foreign currency
for the experimentation results.
Acknowledgment
It is obvious that the development of a book of this scope needs the support of
many people. We must thank Dr. Marianne Caesar, Assistant Development
Editor and the editorial team, IGI Global for their encouragement and support
enabled the book publication project to materialize and contributed to its
success. We especially thank the management of VIT University for their
tremendous assistance. The most important contribution to the development of a
book such as this comes from peer reviews. We cannot express our gratitude in
words to the many reviewers who spent numerous hours reading the manuscript
and providing us with helpful comments and ideas.

We would like to express our sincere gratitude to all the contributors, who have
submitted their high-quality chapters, and to the experts for their supports in
providing insightful review comments and suggestions on time.
 CHAPTER 1
New Computational Models for Image Remote Sensing and Big
Data

Dhanasekaran K. Pillai
Jain College of Engineering, India

ABSTRACT
This chapter focuses on the development of new computational models for
remote sensing applications with big data handling method using image data.
Furthermore, this chapter presents an overview of the process of developing
systems for remote sensing and monitoring. The issues and challenges are
presented to discuss various problems related to the handling of image big data
in wireless sensor networks that have various real-world applications. Moreover,
the possible solutions and future recommendations to address the challenges
have been presented and also this chapter includes discussion of emerging trends
and a conclusion.

INTRODUCTION
The goal of developing new computational models is to enable creation of new
big data based remote sensing infrastructure for analysing and mining image
data. The system must include a data collection component to aggregate,
integrate data and perform validation of image data. Then, the central component
of the system performs tasks like filtering, analysis and extraction of relevant
patterns from image data. The result of extraction and prediction can be used for
agricultural monitoring, crop monitoring or for forecasting of weather and
market values.

Most of the big data framework that uses image remote sensing involves the
following steps:

1. Organizing and integrating data from scenario based models, satellite

 1. Organizing and integrating data from scenario based models, satellite
images, and other remote stations.
2. Developing data mining and correlation analysis techniques to perform
time-series data mining, spatial data analysis or spatiotemporal analysis.
3. Developing classification methods to perform image data classification.
4. Evaluating fitness of the data models by comparing data with standard
values or indices.
5. Developing methods for monitoring activity using images with low or
high resolution.

The system architectural model in Figure 1 involves scenario based models for
analysing and mining images, weather data, and pollution data.

Figure 1. Computational system model using different types of data

For data storage and management, DSpace can be used to store and maintain a
large amount of heterogeneous data. The DSpace is an open source dynamic
digital repository that can be used for image analysis while using big data. It
enables free access to the data.

This chapter enables users to understand major issues and problems related to
remote sensing in combination with big data handling for image data. After
analysing solutions recommended for addressing the problems, users will be able
to understand the process of developing a new framework, tools, or software
systems to meet the current needs.
BACKGROUND
Mostly, remote sensing data is collected to analyse disease conditions, growth of
plant, pollution, land use, road traffic congestions, and effects of disaster etc.
One solution to address these problems is to develop possible computational
models which represent several modules for the data analysis. The creation of
thematic map for certain problems requires meaningful analysis which aims to
show satisfactory results.

The analytical solutions to various engineering problems require robust analysis

in a particular context. In this perspective, various issues related to scenarios
should have to be addressed, because, complexity of different scenarios varies
over time. For most of the existing developments, analysing and developing
computational models have been the main motivation for remote sensing
applications that use images.

The image data collected through multispectral image sensing can provide
information at the element level. It can also provide information at the composite
level via inter-pixel relationships. In some of the applications, the output
information is used to assess the user belief or expert suggestions by the analyst.
The software program validates the hypothesis developed by users. In most of
the cases, the analysis fails due to the compatibility issues between user-defined
performance measure that is used for optimization and the objective that is
unlikely to produce the expected results. So, every analysis should need to be
applied iteratively. The ordering and optimal selection of the objects involved in
analysis may not be known. Hence, an effective approach must use suitable
selection and ordering technique for objects in image remote sensing and
analysis applications.

Mostly, practical engineering problems cannot be solved with hundred percent

perfectness, because, the software system performance largely depends on the
error-tolerance, resource availability, and time taken for the process. Engineering
solutions require experimentation with iterative analysis of algorithms. The new
computational models presented in this chapter will be helpful while proposing a
new system to address various problems related to image remote sensing and big
data.

Crop related mapping of soybean and corn has been conducted at regional scale
focusing on the tropical and temperature plains (Arvor, Jonathan, Meirelles,
focusing on the tropical and temperature plains (Arvor, Jonathan, Meirelles,
Dubreuil, & Durieux, 2011). Most of the methods have used spectral features of
land cover classes for classification either based on supervised learning methods
or based on unsupervised learning methods.

Genetic Programming (GP) uses the principle of natural selection to discover

information using software programs (Alavi & Gandomi, 2011). In fact, GP is a
specialized version of genetic algorithms in which the encoded individual
solutions are software programs rather than binary strings.

GP in some of the recent existing works has focused on the behaviour

characterization and some of the studies have used GP as a tool for interpreting
remote sensing data and can also be used to analyze ground movement patterns,
object movement patterns, and change detection etc.

The crop classes may be identified from phonological information by deriving

phonological metrics and by building classification rules based on crop calendar
and stage-dependent crop conditions (Dong, Xiao, Kou, Qin, Zhang, & Li,
2015).

The effects of temperature, nutrients, and disease have been predicted by

developing an efficient feature classification model (Kuttiyapillai & Rajeswari,
2014). This computational model was focused on information extraction
considering tomato related feature analysis for classification using large margin
k-nearest neighbour classifier.

A context-based method for extracting food safety information was discussed in

a paper titled a method for extracting task-oriented information from biological
text sources (Kuttiyapillai & Rajeswari, 2015). It was focused on information
extraction and dynamic programming technique to find relevant genes from
sequences.

In agricultural field, the detailed information of the planting area is required to

improve the yield estimate, so, it can be incorporated into crop yield models
(Lobell, Thau, Seifert, Engle, & Little, 2015). To find insights into planting area
related and crop related information, image remote sensing that involves
computational models would be a viable option.

Another approach based on machine learning has been developed to select and
combine feature groups. It allows users to give positive and negative examples.
This method improves the user interaction and the quality of queries (Minka, &
Picard, 1997). The two methods discussed in the following paragraphs are based
Picard, 1997). The two methods discussed in the following paragraphs are based
on the concepts of information mining.

Some of the existing approaches does not adapt to different situations to satisfy
user needs. So, the image retrieval has been developed based on relevance
feedback functions (Rui, Huang, Ortega, & Mehrota, 1998). Also, the system is
designed to search image according to the suggestions of user, taking the
feedback into account.

During the last decades, traditional database system has been used to store data
considering characteristics such as color, texture, and shape. Further
development has focused on region-based image retrieval since the content-
based image retrieval was not satisfactory due to the growing size of image and
information content (Veltkamp, Burkhardt, & Kriegel, 2001). This method has
been found to be a viable solution to deal with the varying nature of image
content. In this method, each image is segmented, and object characteristics are
used to index individual object.

Due to the atmospheric effects such as Rayleigh scattering that occurs because of
atmospheric molecules, ozone, absorption by water vapour, and other gases,
absorption due to atmospheric aerosols, changes will frequently occur in the data
collection environment. So, the data correction becomes a computationally
intensive task which requires innovative error correction approaches, for
example, standard radiative transfer algorithms 6S (Vermote, Tanre, Deuze,
Herman, & Morcrette, 1997) for processing high resolution image datasets.
Otherwise, the data processing will not be practically feasible.

MAIN FOCUS OF THE CHAPTER

The advances in remote sensing technologies and availability of Internet access
have increased the amount of data gathered from different locations via satellite
imaging technique. The remote sensing data gathered by a single data center of
satellite are dramatically increasing at several terabytes rate per day. Recent
study shows that the data gathered would even exceed Exabyte range because of
the emergence of high-resolution earth observational data which has led to the
curse of dimensionality issues in terms of image data. These data in some sense
termed as big data.

The way earth system changes would change the accuracy and updation of
The way earth system changes would change the accuracy and updation of
current global earth data. To handle the increasing complexity in terms of size
and dynamism, multi-sensor remote sensing data and temporal big data mining
are useful for big data processing. The major data-intensive computing issues
and challenges arise because of rapid growth of remote sensing data. So, the
methods to deal with the computational complexity associated with big data are
necessarily moving towards the high-performance computing paradigm.

Because of the data availability requirement and huge required computing power
for processing massive amount of data, the cluster-based high-performance
computing still remains as big challenge in remote sensing applications.
Moreover, the intensive irregular data access patterns of remote sensing data
increase I/O burden making the common parallel file systems inapplicable.
Further, the current task scheduling seeks load balancing among computational
resources taking the data availability as major concern. To reduce the
complexity, the large tasks could be divided into smaller data dependent tasks
with ordering constraints. Another way to handling the critical issue is to
introduce an optimization technique for scheduling of tasks while trying to
achieve higher performance.

In addition, specifically the Message Passing Interface (MPI) enabled cluster

systems that support multilevel hierarchy with increasing scale, lead to cause
difficulty and error-prone tasks while using parallel programming techniques for
remote sensing applications. In the near future, the increasing demand for real-
time processing would further increase the issues associated with remote sensing
applications. However, unlike traditional systems, the emerging remote sensing
systems can provide timely real-time processing of sensed data gathered from
environment.

Remote sensing is defined as the technology for perceiving an object or surface

from a remote distance to identify the characteristics of an object. It requires data
acquisition instruments or devices. Remote sensing data are treated as big data
because big data is not only refers to the volume and velocity of data which
causes computational complexity, other issues such as variety, complexity (or
high dimensionality) and trustworthiness also makes these data as big data.

Normally, the processing of remote sensing big data involves the following
stages on the process flow: satellite observational network, data acquisition and
recording, remote sensing data processing (pre-processing, central processing,
information abstraction and representation), and operating remote sensing
applications.
applications.

The pre-processing techniques are used to remove noise and to correct

inconsistent data. It involves techniques like smoothing, outlier detection, and
dimensionality reduction and so on. The central processing component is
responsible for further correction, modification and merging of relevant data.
The information abstraction and representation would involve machine learning
techniques to perform efficient tasks like classification, clustering, and
association analysis.

Some of the examples of remote sensing and monitoring include: disaster

response and monitoring, climate change control and monitoring, water
management and usage monitoring, industry operation control and monitoring,
fraud detection and monitoring etc. Example sensor support includes spatial
sensors, temporal sensors, temperature sensors, humidity sensor, moisture sensor
etc. Example data on remote sensing and monitoring include GPS data, GIS
data, and other sensors’ data. The processing of massive remote sensing data
poses challenging issues; some of these issues are:

1. Difficulty in managing remote sensing big data.

Naturally, the remote sensing data captured from different data centres are
distributed and these data centres are normally far away and connected by the
Internet. So, the data management component has to critically manage these
distributed, huge amounts of data for improving interoperability and global data
sharing.

This means that there is a demand for introducing more storage devices and for
improving easy access. In order to tackle these challenges, new technologies and
techniques are required. Further, the high dimensionality characteristics of
remotely sensed data makes the distributed data sharing and accessing more
complicated. The main issue arises while trying to organize and map the multi-
dimensional remote sensing imageries to a one-dimensional array.

In addition, the remote sensing data partition requires suitable space-filling

curves and efficient data organization depends upon data availability. Moreover,
the complex metadata in structured form associated with the remote sensing data
makes data processing little more complicated. It also affects the efficiency on
data storing and indexing of metadata.

2. The irregular data access pattern.

 2. The irregular data access pattern.

The critical data processing and data sharing component requires high data
availability. Most of the remote sensing applications perform processing on
irregular data access pattern. Example issues include: irregular I/O patterns and
increased CPU load that occurs by varying degree of dependency between
computation of algorithm and remote sensing data.

3. The data loading and transmission of big data.

Because of the high dimensionality of the massive amount of remotely sensed

big data, data loading, memory management, and data transmission becomes
little more complicated and inefficient. The capacity of image big data may
sometimes go beyond the limited capacity of computer memory when multi-
dimensional images are considered along with complex metadata.

Hence, it requires large, efficient data structure to store these massive amounts
of remotely sensed big data in local memory. Further, data transmission among
processing nodes requires high bandwidth to transmit image big data in the form
of data blocks, so, it would be a time-consuming process when the volume of
data to be communicated is large.

4. Scheduling of large number of data-dependent tasks.

The large scale design modeling of water management and remote monitoring
would involve a large number of smaller data dependent tasks. Therefore, the
main processing module become extremely difficult and may require ordering
constraints to deal with data dependent tasks. To achieve good performance, an
optimized scheduling algorithm is required. Sometimes, decoupling of data
dependencies may be helpful to achieve better selection of execution path.

5. Efficient parallel programming.

The conventional clustering systems must deal with a multilevel hierarchical

organization and rapidly increasing scalability issues. It requires efficient
parallel programming techniques for dealing with data intensive computing tasks
in remote sensing and real-time monitoring applications. Recently, OpenMP has
been designed as a new paradigm for shared-memory cluster computing.
SOLUTIONS AND RECOMMENDATIONS
Solutions and recommendations in dealing with the issues, controversies, or
problems are presented in this section. To address various issues and problems
related to remote image sensing, this chapter presents solutions in the form of
computational models and recommends some problem solving techniques.

1. High-Performance Computing Model

The remote sensing and monitoring that requires complex models to work with
large amounts of multi-sensor and temporal data sets may have to apply widely
used pre-processing algorithm. The computational and storage requirements for
problems that use large number of earth observations would normally exceed the
available computing power on a single computing platform.

To deal with this computational complexity, a new computational model is

introduced combining scalable and distributed heterogeneous network with high
performance software program for huge data processing, indexing, and
organizing remote sensing data. The high-performance computing model in
Figure 2 uses a hierarchical data management to achieve load balancing among
the logical sensor processing nodes and also focuses on minimizing I/O
overheads.

Figure 2. High performance computing model for remote image sensing

This computational model deals with error correction on remotely sensed image
data, and includes major components such as image segmentation using
hierarchically connected components, followed by retrieval of data distribution
using computing function, hierarchy based data organization that allows on-
demand processing and retrieval of information. This scheme with addition or
removal of some software modules can be adopted to improve computational
time in remote sensing based applications.
time in remote sensing based applications.

Most of the information analytical systems require data fusion from various
sources and instruments. The use of these data sets in modeling ecosystem
response, various types of data face challenges in dealing with huge data storage
and high computational complexity. Therefore, the data acquisition, processing,
mapping and conversion of remote sensing data deal with a complicated
modeling and increased computational complexity. The computational tasks that
involve a variety of pre-processing of satellite data (e.g. land use data) include
complex neighbourhood operations.

For example, the total storage requirements for a global land cover data sets
would sometimes require processing of Tera bytes, Giga Floating Point
Operations (GFLOPs), or Peta FLOPs for data processing. So, it requires high
performance computing techniques to acquire information that is required from
earth observational systems.

Although innovative parallel processing algorithms have been developed for

processing of large datasets in a heterogeneous, distributed environment, still
there are promising avenues for further improvements and new developments.
Normally, remote image processing includes steps such as pre-processing, image
segmentation, retrieval of distribution via function, and software module for
processing, storing and retrieval of data.

2. Cloud Based Big Data Processing Model

Finding task-relevant information has become a tedious task because of massive

volume of real-time data. To extract meaningful information, the system requires
efficient data analysis, aggregation and storage mechanism while collecting
remote data. The cloud based image big data processing model in Figure 3
consists of four major units such as Data Acquisition unit for remote sensing
data (DA), Local Pre-processing (LP), Local Storage and Processing (LSP), and
Cloud Storage and Retrieval (CSR).

Figure 3. Cloud based remotely sensed image big data processing model
Cloud data processing has capability to handle scalability issue while dealing
with remotely sensed big data. The system that uses earth observational system
is required to perform local processing to purify raw data which usually has
inconsistent data. Because of size and complexity of big data, the traditional
database management faces difficulties in handling big data.

The major challenges in big data processing include: 1) volume that denotes
large amounts of data generated from remote area; 2) velocity that denotes
frequency and speed at which data has been generated and shared; 3) variety that
denotes diversity of data types of data collected from various sources. A system
that deals with any of these challenges may use smaller subsets to create a result
set through correlation analysis. The major disadvantages of conventional
systems include: 1) difficult data transformation of remotely sensed continuous
stream of data; 2) data collected from remote areas are not in a valid format
which is ready for further data analysis; 3) remote sensor network may generate
vast amounts of raw data.

To deal with various challenges in remotely sensed big data analysis, a system
that incorporates offline data storage and filtering with load balancing sub-
systems can be developed for extracting useful information. The input to the big
data system comes from social networks, satellite imagery, sensor devices, Web
servers, finance data store, and banking data store etc.

In this computational model, the load balancer balances the processing power by
distributing the real-time data to the servers where the base station is processing
data. It can also enhance the efficiency of the system. Data extraction finds
insights into the data model and discovers information to create a structured
view of the data. Here, machine learning techniques are applied to process and
interpret image data for generating maps, and summary results. The system that
interpret image data for generating maps, and summary results. The system that
deals with huge amounts of big data processing can be implemented in
development platforms like Python, R Analytics platform, and Hadoop using
MapReduce.

Parallel Processing Model Using I/O Optimization

The data-intensive application that provides on-demand information requires

efficient indexing schemes and data replacement techniques to achieve the
maximum parallel I/O throughput. These techniques are widely used for
handling spatial and temporal datasets with different resolutions.

The atmospheric effects will vary according to the context based on spatial and
temporal data, and also depends on the wavelength and geometry of
observations. Sometimes decoupling of the effects of components will be useful
in remote sensing applications. The feature selection based error removal will
also be helpful in dealing with erroneous data.

The parallel algorithm based on connected components can efficiently perform

while dealing with complicated, computationally intensive tasks like image
enhancement and segmentation for classification of regions in remotely sensed
images. This means that this type of algorithm can improve accuracy and
processing time.

The error correction method involves two steps. The first step is to estimate
atmospheric properties from the imagery. The second is for retrieving surface
reflectance. The following steps are involved in this method:

• Input image (i.e. Window of pixels)

• Identifying dark targets (i.e. dense green vegetation)
• Setting the reflectance threshold
• Compute mean value of pixels whose reflectance threshold is less than the
set threshold
• Estimate reflectance in Thematic Mappers (TM) for the dark pixels
• Use a lookup table to estimate the aerosol optical thickness in Thematic
Mapper bands. Then, check whether the estimate is larger than or equal to
the set aerosol optical thickness. Otherwise, repeat the previous steps with a
smaller threshold until the first aerosol optical thickness is greater than or
equal to the third aerosol optical thickness.
• Use the exponential relationship between wavelength and aerosol optical
 • Use the exponential relationship between wavelength and aerosol optical
thickness to derive coefficients
• Once the aerosol optical thickness is determined for all the TM bands, use
the lookup table to apply error correction in the central pixel of the window.
• The window is moved by incrementing pixel position until entire image is
covered through repeated execution.

The process steps involved in parallel processing model for image remote
sensing is shown in Figure 4.

Figure 4. The parallel processing model for image remote sensing

In this computational model, the correlation between estimated measurements

and ground or standard measurements are verified. The resultant image after
applying error correction method looks clearer than the original image. Also, the
spatial pattern of aerosol optical thickness maintains consistency in comparison
with the normal image. The experimental setting for exchanging messages
simultaneously among all the processor or processing nodes have to connect all
the processors by an Ethernet and a high-performance switch. The approach that
uses single program multiple data model can be suitable to run the same code on
different parts of the input image in parallel.

To implement parallel processing, the input image is divided into a number of

blocks. Then, the equal-sized blocks are distributed among all the processing
nodes. The linear running time is measured in terms of a function of the number
of nodes. This computational model can be useful to minimize the running time
by an efficient error correction method. Both qualitative and quantitative
by an efficient error correction method. Both qualitative and quantitative
analysis is used to improve reliability of the computational model. The modular
approach of implementation of individual components allows easy modification
to the atmospheric properties. The segmentation of remotely sensed imagery is
used to form clusters which contain similar type of region consisting of pixels.
These regions can be classified into categories for improving prediction
accuracy.

Information Mining Model in Remote Sensing Images

The content-based information mining system focuses on managing and

exploring large volumes of remotely sensed image data. This type of system may
include offline processing as well as an interface to online extraction and
processing. The offline module extracts image features, and attempts to improve
compression, data reduction, generation of index for unsupervised image, and
storing content in the database. Further, the semantic interpretations of the image
content (i.e. user’s interest) are associated with Bayesian network based on index
of content. Since this method obtains only a few training samples, the link
computation can be done via online searching on the complete image repository.
The Figure 5 illustrates information mining model for remote sensing images.

Figure 5. Illustration of information mining model for remote sensing

images

During the past few decades, the image satellite sensors such as optical sensors,
synthetic aperture radar, and other satellite sensors have acquired huge amounts
of image scenes. In future, the quantities of real-time image sensing data will
further increase due to the data collection by high resolution satellite sensors.
The state-of-the-art systems access these data and images through query passed
using geographical coordinates, time of acquisition, and sensor type. The
information which are collected using traditional system is often less relevant,
so, only a few images can be used. Because the relevancy is determined by the
so, only a few images can be used. Because the relevancy is determined by the
content of the image considering its structures, patterns, objects, or scattering
properties.

In this approach, a system which incorporates the traditional database concept

for offline images and online search with semantic interpretation are
recommended to develop efficient information mining system that applies
Bayesian network for efficient reasoning. This system can make decision
according to the interest of users. Based on the semantic interpretation of image
content that is linked with index of content in Bayesian network, the user can
search for relevant images which are stored in online repository. Then, the
classification of images is performed to represent image contents that supports
predicted target.

Computational Method for Analyzing and Mining Images

Remote sensing plays a very important role in delivering real-time information
on the crop area for improving productivity, and environment. Traditional
mapping of image patterns is difficult due to the high cost of repeated training
data, inconsistency in interpretation, and the difficulty of handling the varying
weather and crop growth.

In this section, an automated approach to map hybrid tomato and original tomato
is discussed for analyzing and mining image patterns. Here, a decision tree
classifier is constructed by using rules that are manually written based on expert
opinions. The automated approach will be more advantageous when mapping is
to be done for multiple years, because, the mapping can be performed without
re-training or repeated calibration task. To identify vegetation, moderate
resolution, time series based, imaging Spectroradiometer and reflectance product
can be used which identifies hybrid variety of tomato and original variety based
on year.

The surface reflectance of the shortwave infrared band is scaled to identify

similarity between phenology of the two crops. The mapped areas can be
verified with the standard statistical values that are maintained at the municipal
level. The effect of mixed pixel can also be identified by applying classification
technique and the reference dataset can be used to evaluate the resultant map.
Furthermore, the automated mapping can be applied to other image series in
future considering different scales and high-resolution images.

In this model, the following variables are used for measurement and
In this model, the following variables are used for measurement and
classification:

Vb: This denotes the Enhanced Vegetation index value of background in the
non-growing season.
Va: This denotes the amplitude of Enhanced Vegetation index variation
within the growing cycle. The Va value of field crops is higher than natural
vegetation. Average Va of hybrid tomato is greater than original crop.
p, q: changing rate parameters that corresponds to the increased and
decreased segments in the cycle. In this case, the field crop cycles have fast
increase or decrease in enhanced vegetation index value.
Di, Dd: These variables denote the middle Dates of segments when
increasing or decreasing rates are high. These variables are used as
indicators of the dates on which rapid growth is there and harvesting is
necessary.
D1, D2, D3, D4: These variables denote when the second derivative of the
curve reaches local maximum or minimum. D1 denotes starting dates and
D4 denotes end dates of the growing season.
L: This variable denotes the difference between D4 and D1 . It represents
the length of growing season. The length should maintain consistency.
Hybrid tomato has shorter growing time than original tomato.
R: This variable denotes reflectance at Di. Reflectance of Hybrid variety is
slightly higher than normal variety. Here, pixels with high reflectance are
hybrid, and pixels with low reflectance are normal variety and this provides
confidence pixels for training.

Here, the automated mapping is performed based on classification rules in a tree-

like structure. The variables which represent the changes in the vegetation and
stages of transition during the growth period are useful in classification of season
based crop cycles. The Figure 6 illustrates the data and control flow
representation of mapping crop image patterns.

Figure 6. Illustration of the data and control flow representation of

mapping crop image patterns
If there is high variability on seasons, less restrictive rules can be used to
improve the repeatability of algorithm under different conditions. Further,
additional classification rules are necessary based on pheno-spectral variables
that utilize spectral properties at different stages for improving the separation.

Machine Learning Methods for Remote Sensing

Machine learning is a subfield of artificial intelligence. With an intention to

develop algorithm that learns from machine readable data, machine learning has
been evolved as a key technique in many application domains such as data
mining, telecommunication, market analysis, and other software applications. It
includes algorithms such as decision tree, Bayesian reasoning, Naive Bayes, k-
nearest neighbour, support vector machine, neural networks, self-organizing
map, and ensemble methods such as random forests, genetic algorithm, case-
based reasoning, genetic programming, fuzzy logic, neuro-fuzzy, etc. The main
motivation to apply machine learning technique for image remote sensing is to
provide multivariate, nonlinear, and nonparametric regression or classification. It
plays crucial role in science and engineering. Machine learning is capable of
tackling various problems associated with geosciences and remote sensing.

Machine learning has two categories of methods, namely, supervised,

unsupervised, which is used to develop a regression approach or classification
approach of nonlinear systems. The system that involves this technique would
involve a few or literally thousands of variables. In machine learning, training
dataset consists of pre-classified vectors of image data. A subset of the training
samples may not require separate validation. If theoretical knowledge is
incomplete, and there are some observations and data, then, machine learning
would be helpful to address such type of problems.

In the past few decades, machine learning has been a widely used technique in
earth science and observational systems such as land use detection, ocean change
detection, road extraction, and atmosphere, disaster prediction, crop disease
prediction, activity detection, etc. Herein, a number of relevant applications of
machine learning are summarized for understanding its applicability in
geosciences and remote sensing. The main focus may fall under two categories,
one is on how to apply multivariate nonlinear nonparametric regression, and the
other is on how to use multivariate nonlinear unsupervised classification.

The machine learning is actually a universal approximation technique that learns

the underlying behaviour pattern from a set of training data. Another fact about
this technique is that, it does not require prior knowledge about the nature of the
relationships among the data. The application of machine learning includes five
areas: 1) deterministic model which is computationally expensive; 2) non-
deterministic model, or an empirical model that can be derived from the existing
data; 3) classification model; 4) clustering model and; 5) hybrid model. The
study of the successful application of fuzzy logic, neuro-fuzzy, fuzzy-genetic,
neuro-computing can be found in the fields such as oil exploration, intelligent
reservoir exploration, crop area mapping etc.

Among all other machine learning methods, application of genetic programming

(GP) in the field of remote sensing and image analysis has been restricted to a
very few areas and is new also. Although techniques like support vector machine
(SVM) and artificial neural networks (ANNs) have shown good performance,
these machine learning follows black-box models. That is, they are not capable
of producing practical prediction. Hence, GP is considered as a viable technique
of producing practical prediction. Hence, GP is considered as a viable technique
to deal with practical issues in remote sensing applications.

The variants of GP, namely, multi expression programming (MEP), linear

genetic programming (LGP), and gene expression programming (GEP) are used
to predict the uniaxial compressive strength and tensile strength of chalky and
clayey soft limestone. GP models give high prediction than other existing
models, which have used parameters as predictor variables, for example,
modulus of elasticity of intact rock, uniaxial compressive strength, the number
of joint per meter, rock mass quality designation, dry density, and geological
strength index. In rock structure designing, GP can estimate uniaxial
compressive strength of rocks that can be formulated in terms of effective water
absorption by weight, porosity, and actual weight.

Some example applications of GP in remote sensing and big data analysis

include estimation of the heavy rainfall near disaster-prone area which uses
multivariable meteorological satellite data, monitoring water quality using
images, mapping of chemical contaminants in food items, landslide detection
using image threshold, mapping crop planting area and soil moisture analysis.
The major processes in image remote sensing involve image characterization and
image classification.

Image classification is the process of assigning images into a predefined target

classes according to the characteristics of images. It includes techniques like pre-
processing, feature extraction, object detection, object segmentation, and object
classification. Image classification is a challenging task in application domains
such as video surveillance, biomedical imaging, vehicle traffic detection and
navigation, industrial monitoring, remote sensing and real-time monitoring.

Some of the popular machine learning algorithms that can be used in image
classification includes Support Vector Machine, k-nearest neighbour algorithm,
Artificial Neural Network, Genetic Algorithm, Expectation Maximization
Algorithm, C4.5 decision tree algorithm, AdaBoost, CART, k-means algorithm
etc. Mostly, the machine learning algorithms are used to predict the future trends
or quality decisions. Data mining task which involves machine learning aims to
discover information and generate a large number of rules. In big data mining,
image classification mainly focuses on classifying new objects or unknown
vectors under a predefined category or target label.

Image mining is one of the domains which can be used to extract meaningful
image content from a large image dataset. Image classification automatically
image content from a large image dataset. Image classification automatically
classifies image pixels into appropriate target based on natural evaluation and
relationships among image data. It has two categories of classification, namely,
supervised classification, and unsupervised classification.

In disaster response domain, aerial imagery is captured through unmanned

vehicles within few hours whereas satellite imagery can be captured in days.
Also, the spatial resolution of the aerial imagery is higher than the imagery
generated by satellites. To process large volumes of aerial image data in real-
time, machine learning is applied to provide solution. Crowd sourcing provides a
way to annotate features of interest such as damaged shelters, blocked roads by
debris, collapsed building etc. These features of aerial images have been used to
train a supervised machine learning systems which learns to interpret features
from unknown or new images. In addition, the disaster response system includes
a text processing module that can be used to send messages to the rescue team.
Another approach to solve this problem is to develop a hybrid system which
deals with both aerial images as well as satellite images.

FUTURE RESEARCH DIRECTIONS

Remote sensing of image big data for analysis and interpretation has several
special characteristics such as dynamic behaviour, multi-scale, multi-source,
high-dimensional, isomer, and nonlinear characteristics. While analysing which
are the characteristics is closely related to the data acquisition, it is observed that
the intrinsic characteristics include dynamic-state, multi-scale, and nonlinear
characteristics. However, other characteristics are extrinsic characteristics for
remote sensing big data.

Remote sensing provides a method to quickly and directly acquire data from
earth surface. Emerging trends in remote sensing of big data in information
science and environmental engineering has led to the application of remote
sensing and monitoring techniques in various fields which include ecology, earth
quake prediction and analysis, soil contamination, air pollution analysis, water
pollution analysis, environmental geology, solid waste detection and monitoring,
street light monitoring, crop disease analysis and loss prediction, industrial fraud
detection and monitoring, weather forecasting, customer behaviour prediction,
patient monitoring, home appliances monitoring, gas leakage detection and
monitoring in industry, energy consumption and sustainable development etc.
The computational model discussed in this chapter shows the importance of a
particular model in a problem domain.

In recent years, major countries have launched remote sensing satellites, which
include India, USA, and Russia. The remotely sensed features and data may
differ based on image resolution, spectrum, mode of imaging, and revisit cycle,
amplitude, and time. Nowadays, there are different remote sensing systems.
Example low-resolution satellite imaging includes meteorological satellite
MODIS, and microwave satellite Envisat. Example mid-resolution satellite
imaging includes terrestrial satellite (e.g. Landsat), satellite with long revisit
period (e.g. EO-1), microwave satellites (e.g. Terra, and RADARSAT). Example
high-resolution satellite imaging includes QuickBird, IKONOS, and WorldView.
It requires efficient investigations and techniques to deal with increasing
diversity of data.

The satellites can also be classified according to the revisit cycle in an

observational area. For example, Geostationary Operational Environmental
Satellite (GOES) can provide high-quality, continuous stream of time correlated
observational data from Earth surface. Example satellites that use short revisit
cycle include MODIS, WorldView and RapidEye.

Another challenge in handling remote sensing data is to deal with increasing size
or large volume of data. In image remote sensing, for a single scene, the volume
of data may be at the gigabyte level or at the terabyte level. The satellite remote
sensing data collected for a particular period in one country is maintained as
historical data. The volume of these data may be at the petabyte level and a large
archive maintained at the global level may reach up to Exabyte level. Therefore,
remote sensing data is termed as “big data” and requires efficient big data
handling techniques.

CONCLUSION
This chapter presented different computational models for image remote sensing
and big data handling. The representation of remotely sensed image information
on a hierarchical form or in suitable form with different semantic abstraction is
based on the levels involved in computational models. For example, a Bayesian
model may consist of the following levels: 1) extracting image features and
meta-features using signal models; 2) obtaining a vocabulary of signal classes
for each model by applying unsupervised machine learning (or clustering) of the
for each model by applying unsupervised machine learning (or clustering) of the
pre-extracted image parameters; 3) At last, user-interests, i.e., semantic labels are
linked to combinations of these vocabularies through Bayesian networks. In
order to infer information from the image data that covers the class label or
target label, the system has to learn the probabilistic link based on user given
input samples.

Automatic extraction of meaningful information from remote sensing images

involves information mining techniques and robustness evaluation on the
extracted information from the observed image data. This model focused on
extracting structural information from images by selecting prior models that
correctly explains the structures within an image. On the lowest level, stochastic
models are applied to capture spatial, spectral, and geometrical structure
information from image. In recent years, this type of parametric models has been
found to be suitable to characterize spatial information in images. High-order
models are mostly used to deal with complex structures that have features at
different scales (e.g. mountains, rivers, etc.). When neighbourhood size
increases, the number of parameters increases and has to face averaging effect of
different parameters. As a result, less discriminative power of the extracted
features may cause ineffective interpretation of data.

To overcome this drawback, a multi-resolution image data cube that has the
original image at the lowest layer and reduced resolution representations of the
image at the higher layers of data cube may be generated. By applying texture
model that uses Gibbs random field to layers of limited neighbourhood size,
various information from different structures can be extracted. It provides a way
to characterize a large set of spatial information. The output of feature extraction
may have large volumes of data, which may be difficult to store and manage in
practical applications. Moreover, clustering may reduce the accuracy due to a
large data reduction. To remove unnecessary structures and to avoid the time-
consuming process of similarity checking, clustering is performed across all
images. Even if a model generates large number of clusters, the algorithm can
show good efficiency by applying some of these computational models that
focused on parallel processing to improve efficiency of the system to produce
promising result.

REFERENCES
Alavi, A. H., & Gandomi, A. H. (2011). A robust data mining
approach for formulation of geotechnical engineering systems.
Engineering Computations , 28(3), 242–274.
doi:10.1108/02644401111118132

Arvor, D., Jonathan, M., Meirelles, M. S. P., Dubreuil, V., & Durieux,
L. (2011). Classification of MODIS EVI time series for crop mapping
in the state of MatoGrosso. Brazil International Journal of Remote
Sensing , 32(22), 7847–7871. doi:10.1080/01431161.2010.531783

Dong, J., Xiao, X., Kou, W., Qin, Y., Zhang, G., Li, L., & Moore, B.
III. (2015). Tracking the dynamics of paddy rice planting area in
1986-2010 through time series Landsat images and phenology-based
algorithms. Remote Sensing of Environment , 160, 99–113.
doi:10.1016/j.rse.2015.01.004

Kuttiyapillai, D., & Rajeswari, R. (2015). A method for extracting

task-oriented information from biological text sources. International
Journal of Data Mining and Bioinformatics , 12(4), 387–399.
doi:10.1504/IJDMB.2015.070072

Kuttiyapillai, D., & Ramachandran, R. 2014. Design and analysis of

feature classification model using information extraction in tomato
growing environment. International Information Institute (Tokyo).
Information, 17(8), 3947-3959.

Lobell, D. B., Thau, D., Seifert, C., Engle, E., & Little, B. (2015). A
scalable satellite-based crop yield mapper. Remote Sensing of
Environment , 164, 324–333. doi:10.1016/j.rse.2015.04.021

Minka, T. P., & Picard, R. W. (1997). Interactive learning using a

“society of models”. Pattern Recognition , 30(4), 565–581.
doi:10.1016/S0031-3203(96)00113-6
Rui, Y., Huang, T. S., Ortega, M., & Mehrota, S. (1998). Relevance
feedback: A power tool for interactive content-based image retrieval.
IEEE Transaction on Circuits System and Video Technnology , 8(5),
644–655. doi:10.1109/76.718510

Veltkamp, C. R., Burkhardt, H., & Kriegel, H. P. (2001). State-of-the-

Art in content-based image and video retrieval . Norwell, MA:
Kluwer. doi:10.1007/978-94-015-9664-0

Vermote, E. F., Tanre, D., Deuze, J. L., Herman, M., & Morcrette, J.
(1997). Second simulation of the satellite signal in the solar spectrum,
6S: An overview. IEEE Transactions on Geoscience and Remote
Sensing , 35(3), 675–686. doi:10.1109/36.581987

KEY TERMS AND DEFINITIONS

Big Data: The large volume of data, or the complex data that are normally
difficult to handle through simple computational model and techniques.

Computational Model: An abstract representation of the system that performs

computations based on an approach developed by developers or scientists to
solve computational problems.

Data Mining: The process of extracting meaningful information from datasets

that are generated from data sources.

Image Classification: The process of classifying unknown data or new patterns

or new images according to the existing, pre-classified vectors.

Machine Learning: The process of using the existing knowledge and

considering the feedback as the system changes the state to discover required
information in order to produce good results.

Parallel Processing: The process of performing multiple tasks simultaneously,

reducing the waiting time of the process.

Remote Sensing: The task of identifying images and structural characteristics.

Satellite Imaging: The process of using the images from environment, to make
the images in a system-dependent format.
 CHAPTER 2
Big Data Computation Model for Landslide Risk Analysis Using
Remote Sensing Data

Venkatesan M.
National Institute of Technology Karanataka, India

Prabhavathy P.
VIT University, India

ABSTRACT
Effective and efficient strategies to acquire, manage, and analyze data leads to
better decision making and competitive advantage. The development of cloud
computing and the big data era brings up challenges to traditional data mining
algorithms. The processing capacity, architecture, and algorithms of traditional
database systems are not coping with big data analysis. Big data are now rapidly
growing in all science and engineering domains, including biological,
biomedical sciences, and disaster management. The characteristics of complexity
formulate an extreme challenge for discovering useful knowledge from the big
data. Spatial data is complex big data. The aim of this chapter is to propose a
multi-ranking decision tree big data approach to handle complex spatial
landslide data. The proposed classifier performance is validated with massive
real-time dataset. The results indicate that the classifier exhibits both time
efficiency and scalability.

INTRODUCTION
Very large amount of Geo-spatial data leads to definition of complex
relationship, which creates challenges in today data mining research. Current
scientific advancement has led to a flood of data from distinctive domains such
as healthcare and scientific sensors, user-generated data, Internet and disaster
management. Big data is data that exceeds the processing capacity of
conventional database systems. The data is too big, moves too fast, or doesn’t fit
conventional database systems. The data is too big, moves too fast, or doesn’t fit
the strictures of your database architectures. For instance, big data is commonly
unstructured and require more real-time analysis. This development calls forms
system architectures for data acquisition, transmission, storage, and large-scale
data processing mechanisms. Hadoop is a platform for distributing computing
problems across a number of servers. First developed and released as open
source by Yahoo, it implements the MapReduce approach pioneered by Google
in compiling its search indexes. Hadoop’s MapReduce involves distributing a
dataset among multiple servers and operating on the data: the “map” stage. The
partial results are then recombined: the “reduce” stage. To store data, Hadoop
utilizes its own distributed file system, HDFS, which makes data available to
multiple computing nodes.

Natural disasters like hurricanes, earthquakes, erosion, tsunamis and landslides

cause countless deaths and fearsome damage to infrastructure and the
environment. Landslide is the one of the major problem in hilly areas. Landslide
Risk can be identified using different methods based on the GIS technology. In
Ooty, Nilgiri district, landslide was happened due to the heavy rainfall and
frequent modification of land use features. Landslide disaster could have been
reduced, if more had been known about forecasting and mitigation. So far, few
attempts have been made to predict these landslides or prevent the damages
caused by them. In the previous studies, various approaches were applied to such
problems which show that it is difficult to understand and tricky to predict
accurately. In order to analyze these landslides, various factors, such as Rainfall,
Geology, Slope, landuse/land cover, soil and Geomorphology are considered and
the relevant thematic layers are prepared in GIS for landslide susceptibility
mapping. The data collected from various research institutes related to land slide
helped to predict and analyze the land slide susceptibility. The spatial landslide
data is one of the complex big data. To handle such as large amount of landslide
data, the previous study weighted decision tree approach is improvised and Multi
Ranking Decision Tree Classifier is proposed using map reduce programming
model,.

Related Work

Decision trees are one of the most accepted methods for classification in diverse
data mining applications (H. I. Witten & E. Frank, 2005; M. J. Berry & G. S.
Linoff, 1997) and help the development of decision making(J. R. Quinlan,
1990). One of the well known decision tree algorithms is C4.5 (J. R. Quinlan,
1993; J. R. Quinlan, 1996), an expansion of basic ID3 algorithm(J. R.
1993; J. R. Quinlan, 1996), an expansion of basic ID3 algorithm(J. R.
Quinlan,1986). However, with the growing improvement of cloud computing
(M. Armbrust, A. Fox, R. Griffith, A. D. Joseph, R. Katz, A. Konwinski, G. Lee,
D. Patterson, A. Rabkin, I.Stoica and M. Zaharia, 2010) as well as the big data
challenge (D. Howe, M. Costanzo, P. Fey, T. Gojobori, L. Hannick, W. Hide,
D.P. Hill, R. Kania, M. Schaeffer, S.S., 2008), traditional decision tree
algorithms reveal numerous restrictions. First and foremost, building a decision
tree can be very time consuming when the volume of dataset is extremely big,
and new computing paradigm should be applied for clusters. Second, although
parallel computing(V. Kumar, A. Grama, A. Gupta & G. Karypis, 1994) in
clusters can be leveraged in decision tree based classification algorithms (K. W.
Bowyer, L. O. Hall, T. Moore, N. Chawla & W. P. Kegelmeyer, 2000; J. Shafer,
R. Agrawal & M. Mehta, 1996), the strategy of data distribution should be
optimized so that required data for building one node is localized and mean
while the communication cost to be minimized. Weighted classification are well-
suited for many real-world binary classification problems. Weighted
classification (J.L.Polo, F.Berzal, & J.C.Cubero, 2007) assigns different
importance degrees to different attributes. Many different splitting criteria for
attribute selection have been proposed in the literature and they all tend to
provide similar results (F.Berzal, J.C.Cubero, F.Cuenca, & M.J.Martín-Bautista,
2003).

HACE theorem (Xindong Wu,Xingquan Zhu,Gong-Qing Wu, & WeiDing.S,

2014) has been presented which characterizes the features of the Big Data
revolution, and proposes a Big Data processing model, from the data mining
perspective. This data-driven model involves demand-driven aggregation of
information sources, mining and analysis, user interest modeling, and security
and privacy considerations. We analyze the challenging issues in the data-driven
model and also in the Big Data revolution.

An integration of remote sensing, GIS and Data mining techniques has been
used to predicting the landslide risk. The probabilistic and statistical approaches
were applied for estimating the landslide susceptibility area. Landslide
susceptibility map is reduced the landslide hazard and is used for land cover
planning. The frequency ratio model has better than logistic regression model.
Fuzzy membership functions and factor analysis were used to assess the
landslide susceptibility using various factors. The spatial data were collected and
processed and create a spatial database using GIS and Image processing
techniques. The landslide occurrence factor was identified and processed. Each
factor weight was determined and calculated the training using back-
propagation. Improvised Bayesian Classification approach (Venkatesan M*,
propagation. Improvised Bayesian Classification approach (Venkatesan M*,
Rajawat A S, Arunkumar T, Anbarasi M, & Malarvizhi K, 2014) and decision
tree approach (Venkatesan.M, Arunkumar .Thangavelu, & Prabhavathy.P, 2013)
have been applied to predict the landslide susceptibility in Nilgiris district.

Multi Ranking Decision Tree Classifier

Classification is the process to predict the unknown class label using training
data set. Classification approaches are categorized into Decision Tree, Back
propagation Neural Network, Support Vector machine(SVM), Rule based
Classification and Bayesian Classification. In the present scenario, landslide
analysis study was done by using Neural Network and Bayesian but these
approaches are difficult to understand and tricky to predict. In this chapter, Multi
Ranking Decision Tree Classifier is proposed for landslide Risk Analysis. The
performance of the proposed approach is measured with various parameters.

Decision Tree (DT) approach is used to analyze the data in the form of tree. The
Tree is constructed using the top-down and recursive splitting technique. A tree
structure consists of a root node, internal nodes, and leaf nodes. Ranking
classification techniques give simpler models for the important classes. Ranking
classification assigns different importance degrees to different landslide factor.
In this chapter, rankings are assigned to the different landslide factors in order to
represent the relative importance of each landslide factor. In a distributing
computing environment, the large data sets are handled by an open source
framework called Hadoop. It consists of MapReduce, Hadoop file distribution
system (HDFS) and number of related projects Apache Hive, HBase and
Zookeeper.

The Hadoop Distributed File Systems (HDFS) architecture is illustrated in

Figure 1 NameNode is the master node of HDFS handling metadata, and
DataNode is slave node with data storage in terms of blocks. Similarly, the
Master node of Hadoop MapReduce is called JobTracker, which is in charge of
managing and scheduling several tasks, and the slave node is called
TaskTracker, where Map and Reduce procedures are actually performed.

Figure 1. HDFS Architecture

MapReduce programming model is used for parallel and distributed processing
of large datasets on clusters (Venkatesan M*, Rajawat A S, Arunkumar T,
Anbarasi M, & Malarvizhi K, 2014). There are two basic procedures in
MapReduce: Map and Reduce.

In general, the input and output are both in the form of key/value pairs. Figure 2
shows MapReduce programming model architecture. The input data is divided in
to block in the size of 68MB or 128 MB. The mapper input will be supplied as
key/value paris and it produces the relative output in the form of key/pairs.
Partitioner and combiner are used in between mapper and reducer to perform
sorting and shuffling. The Reducer iterates through the values that are associated
with specific key and produces zero or more outputs.
Figure 2. MapReduce Architecture

The dataset is relatively huge in a big data atmosphere, designing appropriate

data structures for parallel programming is very much important. Three data
structures such as attribute table, count table, hash table are used to build parallel
decision tree classifier. Basic information of attribute a, the row identifier of
instance row id, values of attribute values (a) and class labels of instances c are
stored in attribute table . Count table computes the count of instances with
specific class labels if split by attribute a. That is, two fields are included: class
label c and a count cnt .The last one is hash table, which stores the link
information between tree nodes node id and row id, as well as the link between
parent node node id and its branches.

The traditional data is converted into above three data structure for MapReduce
processing. The algorithm -I, procedure data conversion transforms the instance
record into attribute table with attribute Aj as key, and row id and class label c
as values. Then, REDUCEATTRIBUTE computes the number of instances with
specific class labels if split by attribute Aj, which forms the count table.Note that
hash table is set to null at the beginning of process.

Algorithm -I: Data Conversion

Procedure Map-Attribute (tuple_id,(A1,A2,…A3,C))
 emit(Aj,(row_id,C))
End procedure
Procedure Reduce-Attribute ((Aj,(row_id,C))
emit(Aj,(C,Cnt))
End Procedure

In Decision Tree Classifier, selecting best splitting attribute is important task.

The algorithm - II shows that, mapper performs the computation of information
and split information of Aj. The reducer computes the information gain ratio.
The attribute Aj which has maximum value of GainRatio is selected as splitting
attribute.

Algorithm-II: Splitting Attribute Selection

Procedure Reduce_Population((Aj,(C,Cnt))
emit(Aj,all)
End Procedure
Procedure Map_Computation((Aj,(C,Cnt,all)))
Compute
Compute
Compute
emit
End Procedure
Procedure Reduce_Computation( )
emit
End Procedure

As shown in algorithm - III, the records are read from attribute table with key
value equals to abest and emit the count of class labels.

Algorithm-III: Hash Table Updation

Procedure Map_Update_Count((Abest,(row_id,C)))
Emit(Abest, (C,Cnt’))
End Procedure
Procedure Map_Hash((Abest, row_id))
Compute node_id= hash(Abest )
emit(row_id, node_id)
End Procedure

Algorithm –IV shows the procedure to grow the decision tree by building
linkages between nodes.

Algorithm-IV: Building Tree

Procedure Map((Abest,row_id))
Compute node_id=hash(Abest)
If node_id is same with the old value then
emit(row-id,node_id)
end if
add a new subnode
emit(row_id,node_id,subnode_id)
End Procedure

Ranking classification techniques give simpler models for the important classes.
Ranking classification assigns different importance degrees to different landslide
factor. In this chapter, rankings are assigned to the different landslide factors in
order to represent the relative importance of each landslide factor. Weighted
decision tree classification algorithm is improved as multi ranking decision tree
classifier using map reduce programming model as shown in the above
algorithms. The developed classifier is used to analyze the landslide risk in the
ooty region of Niligiris district. The proposed classifier scalability is improved
and performance is compared with the existing classification methods.

Experiment and Result Analysis

The proposed multi ranking decision tree classifier is implemented in Hdoop

cluster. We have HPC cluster with 6 nodes.We let one of them as HDFS
NameNode and MapReduce JobTracker (i.e., master), and the remaining nodes
act as HDFS DataNode and MapReduce TaskTracker (i.e., slave).The efficiency
of weighted decision tree classification algorithm is theoretically and empirically
proved in our previous study. In this paper, we are

concerned with the time efficiency of parallel version of weighted decision tree
classification algorithm in big data environment. This chapter focuses landslide
risk analysis using big data computational techniques. The needed toposheets
and required maps are collected from the geological survey of India. Many
number of factors causes landslide in the hill region, but four factors are very
The performance of proposed classifier is compared with the weighted decision
tree classifier and decision tree classifier on single node. Figure 4 illustrates the
following observations.

Figure 4. Performance of Multi Ranking Decision Tree Classifier

First, the larger the dataset is, the more time consuming it is to build the normal
decision tree approach. Second, the execution time of weighted decision tree
classification takes more time than proposed parallel weighted decision tree
classifier. The proposed MapReduce based multi ranking classifier algorithm
takes less time the original decision tree as the size of dataset increases.
Therefore, it is proved that the proposed multi ranking decision tree classifier
outperforms the sequential version even on a single node environment.

The scalability of the proposed ranking decision tree classification is also tested
in distributed parallel domain. The scalability evaluation includes two aspects:
(1) performance with different numbers of nodes, and (2) performance with
different size of training datasets.

Figure 5. Performance of Multi Ranking Decision Tree Classifier based on

number of nodes
Figure 5 illustrates the execution time of our proposed weighted decision tree
classification algorithm with different numbers of nodes when the number of
record is 1, 2 and 3 lakhs respectively, We have observe that the overall
execution time decreases when the number of nodes increases. This indicates
that the more nodes are involved for computing increases the efficiency of the
algorithm.

CONCLUSION
Predicting and analyzing disaster is complex task. In this chapter, landslide risk
is analyzed using multi ranking decision tree classifier approach. Disaster
management domain generates huge amount of data. Traditional sequential
decision tree algorithms cannot fit to handle such huge data sets. For example, as
the size of training data grows, the process of building decision trees can be very
time consuming. To solve the above challenges, parallel weighted decision
classifier approach is proposed to improve the scalability of the model. We have
compared the performance of the proposed approach with existing approach with
respect to number of nodes and number of record. The empirical results shows
that the proposed algorithm exhibit both time efficiency and scalability. In future
works, the rainfall induced landslide risk analysis will be studied using big data
computational approaches.

REFERENCES
Armbrust, Fox, Griffith, Joseph, & Katz, Konwinski, … Zaharia.
(2010). A view of cloud computing. Communications of the ACM ,
53(4), 50–58.

Berry, M. J., & Linoff, G. S. (1997). Data mining techniques: For

marketing, sales, and customer support . John Wiley & Sons, Inc.

Berzal, F., Cubero, J. C., Cuenca, F., & Martín-Bautista, M. J. (2003).

On the quest for easy-to-understand splitting rules. Data &
Knowledge Engineering , 44(1), 31–48. doi:10.1016/S0169-
023X(02)00062-9

BowyerK. W.HallL. O.MooreT.ChawlaN.KegelmeyerW. P. (2000), A

parallel decision tree builder for mining very large visualization
datasets. IEEE International Conference on Systems Man, and
Cybernetics, 3, 1888-1893. 10.1109/ICSMC.2000.886388

Howe, D., Costanzo, M., Fey, P., Gojobori, T., Hannick, L., Hide, W.,
& Rhee, S. Y. (2008). Big data: The future of biocuration . Nature ,
455(7209), 47–50. doi:10.1038/455047a

Kumar, V., Grama, A., Gupta, A., & Karypis, G. (1994). Introduction
to parallel computing (Vol. 110). Redwood City:
Benjamin/Cummings.

Polo, J. L., Berzal, F., & Cubero, J. C. (2007). Taking class

importance into account. Lecture Notes in Computer Science , 4413.

Quinlan. (1996). Improved use of continuous attributes in C4.5. arXiv

preprint cs/9603103

Quinlan, J. R. (1986). Induction of decision trees . Machine Learning ,

1(1), 81–106. doi:10.1007/BF00116251
Quinlan, J. R. (1990). Decision trees and decision-making . IEEE
Transactions on Systems, Man, and Cybernetics , 20(2), 339–346.
doi:10.1109/21.52545

Quinlan, J. R. (1993). C4.5: programs for machine learning . Morgan

Kaufmann.

ShaferJ.AgrawalR.MehtaM. (1996). PRINT: A scalable parallel

classifier for data mining. Proc. Int.Conf. Very Large Data Bases.

Venkatesan, , Rajawat, Arunkumar, Anbarasi, & Malarvizhi. (2014).

GIS Based Data Mining Classification Approaches for Landslide
Susceptibility Analysis . International Journal of Applied
Environmental Sciences , 9(5), 2345–2357.

Venkatesan, Arunkumar, Thangavelu, & Prabhavathy. (2013). An

Improved Bayesian Classification Data mining Method for Early
Warning Landslide Susceptibility Model Using GIS. Proceedings of
Seventh International Conference on Bio-Inspired Computing:
Theories and Applications (BIC-TA 2012). Springer.

Witten, H. I., & Frank, E. (2005). Data Mining: Practical machine

learning tools and techniques. Morgan Kaufmann.

Wu, , Zhu, , & Wu, , & Ding. (2014). Data Mining with Big Data .
IEEE Transactions on Knowledge and Data Engineering , 26(1).

Zikopoulos, P., & Eaton, C. (2011). Understanding big data: Analytics

for enterprise class hadoop and streaming data . McGraw-Hill
Osborne Media.
 CHAPTER 3
Modified Support Vector Machine Algorithm to Reduce
Misclassification and Optimizing Time Complexity

Aditya Ashvin Doshi

VIT University, India

Prabu Sevugan
VIT University, India

P. Swarnalatha
VIT University, India

ABSTRACT
A number of methodologies are available in the field of data mining, machine
learning, and pattern recognition for solving classification problems. In past few
years, retrieval and extraction of information from a large amount of data is
growing rapidly. Classification is nothing but a stepwise process of prediction of
responses using some existing data. Some of the existing prediction algorithms
are support vector machine and k-nearest neighbor. But there is always some
drawback of each algorithm depending upon the type of data. To reduce
misclassification, a new methodology of support vector machine is introduced.
Instead of having the hyperplane exactly in middle, the position of hyperplane is
to be change per number of data points of class available near the hyperplane. To
optimize the time consumption for computation of classification algorithm, some
multi-core architecture is used to compute more than one independent module
simultaneously. All this results in reduction in misclassification and faster
computation of class for data point.

INTRODUCTION
These days, numerus organizations are using “big data”, “machine learning”
technologies for data analysis. These are the terms which describes that available
technologies for data analysis. These are the terms which describes that available
data is so complex as well as large so that it becomes distinctly clumsy to work
with existing statistical algorithms which restricts size and type of data. The
existing data mining algorithm usually can be divided in to sub types like,
“associate rule mining”, “classification”, “clustering” (A Survey on Feature
Selection Techniques and Classification Algorithms for Efficient Text
Classification.) Classification technique works with association of unstructured
to well-structured data. Numerous amount of classification techniques is
introduced in the fields of big data. As every algorithm has its own pros and cons
depending upon type of data need to be classified. The performance of these
techniques is generally measured in terminology of cost and cost is nothing but
required computation time and misclassification.

What is Machine Learning?

Consider “machine learning” this way. As a human being, and as a client of

innovation, you finish certain errands that oblige you to choose or group
something. For example, when you read your inbox in the morning, you choose
to stamp that “Win a Free Cruise on the off chance that You Click Here” email
as spam. How might a computer know to do a similar thing? Machine learning is
involved algorithms that instruct computers to perform assignments that
individuals do day by day.

The primary endeavour’s at counterfeit consciousness included instructing a

computer by composing a run the show. On the off chance that we needed to
educate a computer to make proposals considering the climate, then we may
compose a decide that stated: IF the climate is shady AND the possibility of
precipitation is more prominent than half, THEN recommend taking an
umbrella. The issue with this approach utilized as a part of conventional master
frameworks, in any case, is that we don't know how much certainty to put on the
run the show.

Hence, machine learning has developed to imitate the example coordinating that
human brains perform. Today, machine learning algorithms instruct computers
to perceive components of a protest. In these models, for instance, a computer is
demonstrated an apple and told that it is an apple. The computer then uses that
data to characterize the different attributes of an apple, expanding upon new data
each time. At initial, a computer may arrange an apple as round, and fabricate a
model that expresses that if something is around, it's an apple. At that point later,
when an orange is presented, the computer discovers that if something is around
AND red, it's an apple. At that point a tomato is presented, etc. The computer
should persistently alter its model considering new data and allot a prescient
incentive to each model, showing the level of certainty that a question is one
thing over another. For instance, yellow is a more prescient incentive for a
banana than red is for an apple.

So Why is Everyone Talking about Machine Learning?

These essential algorithms for instructing a machine to finish assignments and

order like a human go back a very long while. The contrast amongst now and
when the models were initially designed is that the more data is sustained into
the algorithms, the more precise they move toward becoming. The previous
couple of decades have seen enormous versatility of information and data, taking
into consideration a great deal more precise forecasts than were ever conceivable
in the long history of machine learning.

New systems in the field of machine learning – that generally include

consolidating pieces that as of now existed in the past – have empowered a
remarkable research exertion in “Deep Neural Networks (DNN)”. This has not
been the consequence of a noteworthy leap forward, yet rather of substantially
speedier computers and a great many analysts contributing incremental
upgrades. This has empowered scientists to extend what's conceivable in
machine learning, to the point that machines are beating people for troublesome
yet barely characterized errands, for example, perceiving appearances or playing
the round of Go.

Why is this Important?

Machine learning has a few exceptionally functional applications that drive the
sort of genuine business comes about-- for example, time and cash funds – that
can possibly significantly affect the eventual fate of your association. At
Connections, specifically, we see colossal effect happening inside the client
mind industry, whereby machine learning is permitting individuals to
accomplish things more rapidly and proficiently. Through virtual right-hand
arrangements, machine learning computerizes errands that would some way or
another should be performed by a live operator –, for example, changing a secret
key or checking a record adjust. This arranges for significant operator time that
can be utilized to concentrate on the sort of client care that people perform best:
high touch, confused basic leadership that is not as effectively dealt with by a
machine. At Associations, we additionally enhance the procedure by wiping out
machine. At Associations, we additionally enhance the procedure by wiping out
the choice of whether a demand ought to be sent to a human or a machine: one
of a kind versatile understanding innovation, the machine figures out how to
know about its constraints, and safeguard to people when it has a low trust in
giving the right arrangement.

In “Support Vector Machine”, decision boundaries are defined depending upon

decision plane. A plane that separate data points of different classes (Ali
AlShaari, M., 2014). The example is illustrated below in Figure 1. In this
example, there are data points of two classes which are represented with two
different color. In this figure one side having data points of same class and other
side having data points of another class which is separated by line. New data
points class can be decided depending upon that data point belong to which side
of that separating line. The classification algorithm is of two types “linear
classification” and “non-linear classification”. “Classification” will be linear if
and only if there exist a clear straight line which separates data points into
distinct group, then that is nothing but linear classification (Guang-chao, W.,
2008). If there is no clear straight line separating two different classes, then that
is non-linear classification.

Figure 1. SVM Classifier

The above is a classic example of a linear classifier, i.e., a classifier that
separates group of data points in to their respective category (GREEN and RED
in this case) with a line. Most of the classification work is more complex
structured and making optimal separation is needed and this is not that simple,
i.e., correctly classify new data point based on the existing data points with
known class. This situation is depicted in the illustration below. Compared to the
previous schematic, a full separation of the GREEN and RED objects would
require a curve (which is more complex than a line). Classification tasks based
on drawing separating lines to differentiate among data points of different class
member are known as hyperplane classifiers. “Support Vector Machines” are
particularly suited to handle such tasks.

Figure 2. SVM Classifier

The representation beneath demonstrates the essential thought behind “Support
Vector Machine”. Here we see the original data points are mapped, i.e., adjusted,
utilizing an of mathematical functions, known as kernels. The way toward
reworking the articles is known as mapping (change) (Gumus, F., 2014). Take
note of that in this new setting, the mapped data points are linearly divisible and,
in this manner, rather than building the complex curve Figure 2 (left schematic),
we should simply to locate an ideal line that can isolate the GREEN and the
RED data points.

BACKGROUND
Classification technique works with association of unstructured to well-
structured data. Numerous amount of classification techniques is introduced in
the fields of big data. As every algorithm has its own pros and cons depending
the fields of big data. As every algorithm has its own pros and cons depending
upon type of data need to be classified. The performance of these techniques is
generally measured in terminology of cost and cost is nothing but required
computation time and misclassification.

The Role of Classification in Data Mining

In view of the information gathered, information mining algorithms are utilized

to either create a depiction of the information put away, or anticipate an outcome
(Laskov, P., & Lippmann, R., 2010). Various types of algorithms are utilized to
accomplish both assignments. Nonetheless, in the general KDD prepare, any
blend of these assignments might be called upon to accomplish the coveted
outcomes. The Steps required in KDD are:

1. Description Errands: These assignments depict the information being

mined and they are:
a. Summarization: To concentrate minimal patterns that portray
subsets of information. The technique used to accomplish this
undertaking are Association Rule algorithms.
b. Segmentation or Clustering: To separate information things into
subsets that are like each other. Segment based grouping algorithms
are utilized to accomplish this undertaking.
c. Change and Deviation Detection: To distinguish changes in
successive information, (for example, protein sequencing, behavioral
groupings, and so on.).
d. Dependency Modeling: To build models of causality inside the
information.
2. Prediction Assignments: To foresee some field(s) in a database in light
of data in different fields.
a. Classification: To foresee the in all likelihood condition of a clear
cut variable (its class).
b. Regression: To foresee comes about that are numeric persistent
factors.

Classification

• Model Construction: Model development is building the model from the

preparation set.
o Each tuple/test is expected to have a place a prefined class
o The class of a tuple/test is dictated by the class name quality
 o The class of a tuple/test is dictated by the class name quality
o The preparation set of tuples/tests is utilized for model development
o The model is spoken to as arrangement principles, choice trees or
scientific formulae
• Model Usage
o Group future or obscure items
o Appraise exactness of the model
o The known class of a test tuple/test is contrasted and the outcome
given by the mode
o Accurate rate = rate of the tests tuples/tests effectively arranged by
the model

Existing Methods in Data Classification

With a colossal measure of information put away in databases and information

distribution centres, it is progressively imperative to grow capable apparatuses
for investigation of such information and mining intriguing learning from it.
Information mining is a procedure of construing learning from such tremendous
information. Information Mining has three noteworthy parts Clustering or
Classification, Association Rules and Sequence Analysis. Information
Classification is a critical stride in information mining applications.

By straightforward definition, in characterization/bunching we break down an

arrangement of information and produce an arrangement of collection principles
which can be utilized to group future information. For instance, one may order
infections and give the side effects which portray each class or subclass. This has
much in a similar manner as conventional work in measurements and machine
learning. In any case, there are vital new issues which emerge due to the sheer
size of the information. One of the essential issue in information mining is the
Classification-lead learning which includes discovering decides that parcel given
information into predefined classes. In the information mining space where a
great many records and countless are included, the execution time of existing
algorithms can wind up noticeably restrictive, especially in intuitive
applications. In Data characterization one builds up a portrayal or model for each
class in a database, in view of the components exhibit in an arrangement of
class-named preparing data. There have been numerous information arrangement
techniques, for example, choice tree strategies, for example, statistical
techniques, neural systems, harsh sets, database-situated techniques and so forth.

Data Classification Methods

Data Classification Methods

The accompanying rundown demonstrates the accessible information

arrangement techniques.

• Statistical Algorithms: Statistical examination frameworks, for example,

SAS and SPSS have been utilized by experts to identify surprising patterns
and clarify patterns utilizing statistical models, for example, direct models.
Such frameworks have their place and will keep on being utilized.
• Neural Networks: Artificial neural systems copy the pattern-discovering
limit of the human cerebrum and consequently a few specialists have
recommended applying Neural Network algorithms to pattern mapping.
Neural systems have been connected effectively in a couple of utilizations
that include grouping.
• Genetic Algorithms: Optimization systems that utilization procedures,
for example, hereditary blend, transformation, and normal determination in
an outline in view of the ideas of common development.
• Nearest Neighbour Method: A procedure that orders each record in a
dataset considering a blend of the classes of the k record(s) most like it in a
verifiable dataset. Here and there called the k-closest neighbour method.
• Rule Enlistment: The extraction of helpful if-then principles from
information considering statistical importance.
• Data Visualization: The visual elucidation of complex connections in
multidimensional information.

Many existing methods propose abstracting the test information before arranging
it into different classes. There are a few options for doing reflection before order:
An informational index can be summed up to either a negligibly summed up
deliberation level, a transitional deliberation level, or a high deliberation level.
Too low a deliberation level may bring about scattered classes, thick grouping
trees, and trouble at compact semantic understanding; though too high a level
may bring about the loss of characterization exactness.

1. Genetic Algorithm

Development has turned out to be an effective instrument in discovering great

answers for troublesome issues. One can take a gander at the normal
determination as an advancement strategy, which tries to create sufficient
answers for specific conditions.

Disregarding the expansive number of uses of GA in various sorts of

Disregarding the expansive number of uses of GA in various sorts of
enhancement issues, there is almost no examination on utilizing this sort of way
to deal with the grouping issue. And remembering the nature of the
arrangements that this innovation has appeared in changed sorts of fields and
issues (Beasley, Bull & Martin, 1993a, & Mitchell, 1996) it bodes well to
attempt to utilize it in grouping issues.

The adaptability related with GA is one critical perspective to hold up under as a

top priority. With a similar genome portrayal and just by changing the wellness
work one can have an alternate algorithm. Because spatial investigation this is
especially vital since one can attempt distinctive wellness works in an
exploratory stage.

In the genome, every quality speaks to an information point and characterizes

group enrolment. All vital development administrators can be executed with this
plan. As pointed by Demiriz (1999) each of the real issue related with this
portrayal plan is that it is not adaptable, then again it is by all accounts
computationally effective when the quantity of information focuses is not very
expansive.

2. Decision Trees

A choice tree is a characterization plan which produces a tree and an

arrangement of principles, speaking to the model of various classes from a given
information set. The arrangement of records accessible for creating
characterization strategies is for the most part partitioned into two disjoint
subsets as takes after:

a. A preparing set: Utilized for inferring the classifier.

b. A test set: Used to gauge the precision of the classifier.

The precision of the classifier is dictated by the rate of the test cases that are
accurately arranged. The characteristics of the records are partitioned into two
sorts as takes after:

a. Numerical traits: Qualities whose space is numerical.

b. Clear cut qualities: Traits whose area is not numerical.

There is one recognized trait called the class mark. The objective of the
arrangement is to assemble a compact model that can be utilized to anticipate the
class of the records whose class mark is not known.
class of the records whose class mark is not known.

A choice tree is a tree where the interior hub - is a test on a quality, the tree limb
- is a result of the test, and the leaf hub - is a class name or class dispersion.

There are two periods of choice tree generation:

Tree Construction

• At begin, all the preparation cases are at the root

o Partition cases considering chose qualities
o Test traits are chosen considering a heuristic or a statistical measure

Tree pruning

• Recognize and expel branches that reflect clamour or anomalies

o One govern is produced for every way in the tree from the root to a
leaf
o Each characteristic esteem match along a way structures a
conjunction
o The leaf hub holds the class expectation
o Rules are for the most part less difficult to comprehend than trees

Tree Construction Principle

There are different techniques for building choice trees from a given preparing
informational index. Some essential ideas required in the working of choice trees
are examined beneath.

Splitting Attribute

With each hub of the choice tree, there is a related property whose qualities
decide the apportioning of the informational index when the hub is extended.

Splitting Criterion

The qualifying condition on the part property for informational collection part at
a hub is known as the part foundation at that hub. For a numeric characteristic,
the paradigm can be a condition or an imbalance. For a clear-cut trait, it is a
participation condition on a subset of qualities.
participation condition on a subset of qualities.

3. Decision Tree Construction Algorithms

Various algorithms for instigating choice trees have been proposed throughout
the years. They vary among themselves in the techniques utilized for choosing
part properties and part conditions. These algorithms can be arranged into two
sorts. The principal kind of algorithms is the established algorithms which
handle just memory occupant information. The second class can deal with the
proficiency and adaptability issues. These algorithms evacuate the memory
limitations and are quick and versatile.

4. Naïve K-Implies algorithm

A standout amongst the most prominent heuristics for taking care of the k-means
issue depends on a basic iterative plan for finding a locally ideal arrangement.
This algorithm is regularly called the k-implies algorithm. There are various
variations to this algorithm, so to illuminate which rendition we are utilizing, we
will allude to it as the credulous k-implies algorithm as it is substantially less
complex contrasted with alternate algorithms portrayed here.

The guileless k-implies algorithm segments the dataset into “k” subsets with the
end goal that all records, starting now and into the foreseeable future alluded to
as focuses, in each subset “have a place” to a similar focus. Likewise, the
focuses in each subset are nearer to that inside than to some other focus. The
parcelling of the space can be contrasted with that of Veronesi dividing aside
from that in Veronesi apportioning one segments the space in view of separation
and here we segment the focuses considering separation. The algorithm monitors
the centroids of the subsets, and continues in straightforward emphases. The
underlying dividing is haphazardly created, that is, we arbitrarily instate the
centroids to a few focuses in the locale of the space. In every cycle step, another
arrangement of centroids is created utilizing the current arrangement of centroids
taking after two extremely straightforward strides. Give us a chance to mean the
arrangement of centroids after the ith emphasis by

The accompanying operations are performed in the means:

• Partition the focuses in view of the centroids C (i), that is, discover the
centroids to which each of the focuses in the dataset has a place. The
focuses are apportioned in view of the Euclidean separation from the
centroids.
 centroids.
• Set another centroid c(i+1) ∈ C (i+1) to be the mean of the considerable
number of focuses that are nearest to c(i) ∈ C (i) The new area of the
centroid in a specific segment is alluded to as the new area of the old
centroid.

The algorithm is said to have joined while re-computing the allotments does not
bring about an adjustment in the apportioning. In the wording that we are
utilizing, the algorithm has merged totally when C(i) and C(i – 1) are
indistinguishable. For designs where no point is equidistant to more than one
focus, the above joining condition can simply be come to. This meeting property
alongside its straightforwardness adds to the engaging quality of the k-implies
algorithm.

The naive k-means needs to play out countless “neighbour” inquiries for the
focuses in the dataset. If the information is “d” dimensional and there are “N”
focuses in the dataset, the cost of a solitary emphasis is O(kdN). As one would
need to run a few cycles, it is for the most part not achievable to run the innocent
k-means algorithm for expansive number of focuses.

Here and there the union of the centroids (i.e. C(i) and C(i+1) being
indistinguishable) takes a few cycles. Likewise, in the last a few emphases, the
centroids move practically nothing. As running the costly cycles such many
more circumstances won't not be proficient, we require a measure of joining of
the centroids with the goal that we stop the emphases when the merging criteria
is met. Twisting is the most broadly acknowledged measure.

Bunching mistake measures a similar rule and is occasionally utilized rather than
bending. Truth be told k-means algorithm is intended to improve twisting.
Putting the bunch focus at the mean of the considerable number of focuses limits
the bending for the focuses in the group. Likewise, when another group focus is
more like a point than its present bunch focus, moving the group from its present
bunch to the next can lessen the bending further. The over two stages are
correctly the means done by the k-means group. Along these lines k-means
decreases twisting in each progression locally. The k-Means algorithm ends at an
answer that is locally ideal for the twisting capacity. Thus, a characteristic
decision as a meeting standard is contortion. Among different measures of
merging utilized by different analysts, we can quantify the total of Euclidean
separation of the new centroids from the old centroids. In this proposition, we
generally utilize grouping blunder/twisting as the union rule for all variations of
k-means algorithm.
k-means algorithm.

5. The Greedy K-Means Algorithm

The nearby union properties of k-means have been enhanced in this algorithm.
Likewise, it doesn't require the underlying arrangement of centroids to be
chosen. The thought is that the worldwide minima can be come to through a
progression of neighbourhood ventures in view of the worldwide bunching with
one group less.

Assumption: The presumption utilized as a part of the algorithm is that the

worldwide optima can be come to by running k-means with the (k-1) clusters
being put at the ideal positions for the (k-1) grouping issue and the kth bunch
being set at a fitting position that is yet to be found.

Give us a chance to accept that the issue is to discover K clusters and K' ≤ K. We
use the above suspicion, the worldwide optima for k = K' clusters is registered as
a progression of neighbourhood quests. Expecting that we have tackled the k-
means bunching issue for K' – 1 clusters, we need to put another group at a
fitting area. To find the fitting addition area, which is not known, we run k-
means algorithm until joining with each of the focuses in the whole arrangement
of the focuses in the dataset being included as the applicant new bunch, each one
in turn, to the K' – 1 clusters. The focalized K clusters that have the base
mutilation after the joining of k-means in the above neighbourhood quests are
the clusters of the worldwide k-means. We realize that for k = 1, the ideal
grouping arrangement is the mean of the considerable number of focuses in the
dataset. Utilizing the above technique, we can figure the ideal positions for the k
= 2, 3, 4, ... K, clusters. Subsequently the procedure includes figuring the ideal k-
means communities for each of the K = 1, 2, 3… K clusters. The algorithm is
completely deterministic.

Although the engaging quality of the worldwide k-means lies in it finding the
worldwide arrangement, the technique includes a substantial cost. K-means is
run N times, where N is the quantity of focuses in the dataset, for each group to
be embedded. The unpredictability can be decreased significantly by not running
the K-means with the new group being embedded at each of the dataset focuses
however by finding another arrangement of focuses that could go about as a
proper set for inclusion area of the new bunch.

The variation of the kd-tree parts the focuses in a hub utilizing the plane that
goes through the mean of the focuses in the hub and is opposite to the essential
goes through the mean of the focuses in the hub and is opposite to the essential
segment of the focuses in the hub. A hub is not part if it has not exactly a pre-
indicated number of focuses or an upper bound to the quantity of leaf hubs is
come to. The thought is that regardless of the possibility that the kd-tree were
not utilized for closest neighbour questions, only the development of the kd-tree
in light of this system would give a decent preparatory bunching of the
information. We can accordingly utilize the kd-tree hubs focuses as the
applicant/introductory inclusion positions for the new clusters. The time
multifaceted nature of the algorithm can likewise be enhanced by adopting an
eager strategy. In this approach, running k-means for every conceivable
inclusion position is stayed away from. Rather diminishment in the contortion
when the new group is included is considered without running k-means. The
point that gives the most extreme reduction in the mutilation when included as a
group focus is taken to be the new addition position.

K-means is keep running until union on the new rundown of clusters with this
additional point as the new group. The presumption is that the point that gives
the most extreme abatement in twisting is likewise the point for which the
focalized clusters would have the minimum mutilation. These outcomes in a
considerable change in the running time of the algorithm, as it is pointless to run
k-means for all the conceivable inclusion positions. Be that as it may, the
arrangement may not be all inclusive ideal but rather an inexact worldwide
arrangement.

6. Self-Organizing Map

The SOM can be described as:

an unsupervised system that tries to take in a ceaseless

topological mapping of an arrangement of sources of info
onto an arrangement of yields such that the yields obtain an
indistinguishable topological request from the contributions,
by means of self-association in light of information cases
(Openshaw & Wymer 1994).

Neurons are ordinarily sorted out in a 2D matrix, and the SOM tries to discover
clusters to such an extent that any two clusters that are near each other in the
lattice space have codebook vectors that are near each other in the information
space.

In the self-association prepare the information vectors are introduced to the

In the self-association prepare the information vectors are introduced to the
system, and the group unit whose weight vector is nearest (as a rule as far as
Euclidean separation) is picked as the victor. The following stride is to refresh
the estimation of the triumphant unit and neighbouring units, this will inexact the
estimations of the units to the one of the information vector. This be a movement
of the units toward the info vector, the extent of this development relies on upon
the learning rate, which diminishes along the procedure to get joining.
Remembering that Vector Quantization (VQ) is basically the same as the k-
means algorithm, and that the VQ is an exceptional instance of the SOM, in
which the area size is zero, one can state that there is a nearby connection
amongst SOM and k-means. Openshaw and Wymer (1994) go further and say
that the fundamental SOM algorithm

… is basically the same as a K means classifier; with a couple

of contrasts because of neighbouring preparing which may
well be viewed as a type of recreated tempering and it might
give better outcomes and dodge some nearby optima.

Machine learning techniques can also be divided as supervised leaning and

unsupervised leaning.

Supervised Learning

Supervised learning is an algorithm in which both the information sources and

yields can be seen. Considering this preparation information, the algorithm needs
to sum up with the end goal that it can accurately react to every conceivable info
(Kulesza, A., 2012). This algorithm is relied upon to create amend yield for
sources of info that weren't experienced amid preparing. In supervised learning,
what must be discovered is determined for every illustration. Supervised
arrangement happens when a coach gives the characterization to every
illustration. Supervised learning of activities happens when the operator is given
prompt input about the estimation of each activity (Kulesza, A., 2012). To tackle
a give issue utilizing administered learning algorithm one must take after some
specific strides:

1. Determine the kind of training illustrations.

2. Gather a training set.
3. Determine the info include portrayal of educated capacity.
4. Determine the structure of learning capacity and relating learning
algorithm.
 algorithm.
5. Complete the plan and run the learning algorithm on the assemble set of
information.
6. Evaluate the exactness of the educated capacity likewise the execution of
the learning capacity ought to be measured and after that the execution
ought to be again measured on the set which is not the same as the
preparation set.

Figure 3. Supervised Learning

Unsupervised Learning

Unsupervised learning considers how frameworks can figure out how to speak to
specific information patterns in a way that mirrors the measurable structure of
the general gathering of info patterns (Laskov, P., & Lippmann, R., 2010). By
appear differently in relation to “SUPERVISED LEARNING or
REINFORCEMENT LEARNING”, there are no express target yields or natural
assessments related with each info; rather the unsupervised learner conveys to
hold up under earlier predispositions concerning what parts of the structure of
the information ought to be caught in the yield (Laskov, P., & Lippmann, R.,
2010).

Unsupervised learning is essential since it is probably going to be a great deal

more typical in the mind than supervised learning. For example, there are around
more typical in the mind than supervised learning. For example, there are around
photoreceptors in each eye whose exercises are always showing signs of change
with the visual world and which give all the data that is accessible to
demonstrate what protests there are on the planet, how they are displayed, what
the lighting conditions are, and so forth. Formative and grown-up versatility are
basic in creature vision (see VISION AND LEARNING) – auxiliary and
physiological properties of neurotransmitters in the neocortex are known to be
significantly impacted by the patterns of movement in tangible neurons that
happen. Be that as it may, basically none of the data about the substance of
scenes is accessible amid learning. This makes unsupervised strategies
fundamental, and, similarly, permits them to be utilized as computational models
for synaptic adjustment.

The main things that unsupervised learning techniques need to work with are the
watched input patterns, which are regularly thought to be free specimens from a
basic obscure likelihood conveyance, and some express or certain from the
earlier data in the matter of what is essential. One key idea is that info, for
example, the picture of a scene, has distal autonomous causes, for example,
objects at given areas lit up by specific lighting (Laskov, P., & Lippmann, R.
(2010). Since it is on those free causes that we regularly should act, the best
portrayal for an information is in their terms. Two classes of technique have
been recommended for unsupervised learning. Thickness estimation methods
expressly fabricate factual models, (for example, “BAYESIAN NETWORKS”)
of how hidden causes could make the information. Include extraction systems
attempt to separate measurable regularities (or now and again inconsistencies)
straightforwardly from the information sources (Laskov, P., & Lippmann, R.,
2010).

Unsupervised learning as a rule has a long and recognized history. Some early
impacts were “Horace Barlow” (see Barlow, 1992), who looked for methods for
portraying neural codes, “Donald MacKay” (1956), who embraced a robotic
theoretic approach, and “David Marr” (1970), who made an early unsupervised
learning propose about the objective of learning in his model of the neocortex.
The Hebb administer (Hebb, 1949), which joins measurable techniques to
neurophysiological trials on versatility, has likewise thrown a long shadow.
“Geoffrey Hinton” and “Terrence Sejnowski” in designing a model of learning
called the Boltzmann machine (1986), imported a considerable lot of the ideas
from insights that now command the thickness estimation techniques
(Grenander, 1976-1981). Highlight extraction strategies have been less broadly
investigated.
investigated.

Bunching gives an advantageous case. Consider the case in which the sources of
info are the photoreceptor exercises made by different pictures of an apple or an
orange. In the space of all conceivable exercises, these specific sources of info
shape two bunches, with numerous less degrees of variety than, yellower
measurement. One regular errand for unsupervised learning is to discover and
describe these different, low dimensional bunches.

The bigger class of unsupervised learning techniques comprises of most extreme

probability (ML) thickness estimation strategies. These depend on building
parameterised models (with parameters) of the likelihood dispersion, where the
types of the models (and potentially earlier appropriations over the parameters)
are compelled by from the earlier data as the representational objectives.

The littler class of unsupervised learning techniques looks to find how to speak
to the contributions by characterizing some quality that great 6 features have,
and after that hunting down those elements in the sources of info. For example,
consider the case that the yield is a direct projection of the info onto a weight
vector. As far as possible hypothesis infers that most such straight projections
will have Gaussian insights. Consequently, on the off chance that one can
discover weights with the end goal that the projection has an exceptionally non-
Gaussian (for example, multi-modular) conveyance, then the yield is probably
going to mirror some fascinating part of the info. This is the instinct behind a
factual technique called projection interest. It has been demonstrated that
projection interest can be actualized utilizing an adjusted type of Hebbian
learning (Intrator & Cooper, 1992) (Laskov, P., & Lippmann, R., 2010).
Orchestrating that distinctive yields ought to speak to various parts of the info
ends up being shockingly precarious.

Projection interest can likewise execute a type of grouping in the case. Consider
anticipating the photoreceptor exercises onto the line joining the focuses of the
bunches. The circulation of all exercises will be bimodal – one mode for each
bunch – and thusly exceptionally non-Gaussian. Take note of that this single
projection does not describe well the nature or state of the groups.

Another case of a heuristic fundamental great elements is that causes are

regularly to some degree worldwide. For example, consider the visual
contribution from a protest saw top to bottom. Diverse parts of the protest may
share few components, except for that they are at a similar profundity, i.e. one
part of the divergence in the data from the two eyes at the different areas is
comparative. This is the worldwide hidden component. By boosting the shared
data amongst yields and that are figured on the premise of the different
information, one can discover this difference. This strategy was designed by
Becker and Hinton (1992) and is called IMAX.

Some of the Supervised Learning Methods are:

1. Support Vector Machine

Title “Support Vector Machine” (SVM) is a standout amongst the best

classification algorithms in the data mining area, yet it's long preparing time
constrains its utilization. The point of SVM is to discover ideal separating hyper
plane with maximum margin between the two classes, which offer the good
generalization capacity for future data. Their storage and computation necessities
increment quickly with the quantity of preparing vectors (Ali AlShaari, M.,
2014). SVM utilizes statistical learning hypothesis to boost generalization
property of produced classifier.

Assume some given information indicates each have a place one of two classes,
and the objective is to choose which class another information point will be in.
In “Support vector machines”, data point is a P-dimensional vector, and we need
to know whether we can separate such data points with (P-1)- dimensional
hyperplane (Ali AlShaari, M., 2014). This is known as a linear classifier. There
are numerous hyperplanes that may classify the information. The best choice of
hyperplane is such line with highest margin from the different data points. So,
we pick the hyperplane so that the separation from it to the closest data point on
each side is maximum. Such hyperplane is known as hyperplane with maximum
margin.

2. K Nearest Neighbour

“K Nearest Neighbour (KNN)” is a standout amongst the most generally utilized

classification algorithm in data mining, which is dependent of learning by
relationship, that is by contrasting a given test tuple and existing tuples which
resemble it (Dayan, P). K-NN classification decides the decision boundary
locally that was produced from the need to perform segregate investigation when
dependable parametric assessments of probabilistic densities are obscure or hard
to decide.
Figure 4. KNN Classifier

The existing data cases are vectors in a multidimensional element space, each
with a class mark. The training period of the algorithm comprises just of putting
away the component vectors and class names of the existing data points.

In the classification stage, k is a constant which is determined by user depending

upon type of data, and a non-labeled vector is grouped by appointing the name
which is most regular among the k preparing tests closest to that required data
point.

An ordinarily utilized metric for continuous data is “Euclidean distance”. For

discrete data, for example, for content classification, another metric can be
utilized, for example, the overlapping metric (or “Hamming distance”) (Dubey,
V., 2016). With regards to quality expression microarray data, for instance, k-
NN has likewise been utilized with connection coefficients, for example,
Pearson and Spearman. Frequently, the classification precision of k-NN can be
enhanced fundamentally if the distance metric is found out with algorithms.

MAIN FOCUS OF THE CHAPTER

The “Support Vector Machine” (SVM) is a standout amongst the best
classification algorithms in the data mining area, yet it's long preparing time
constrains its utilization. The point of SVM is to discover ideal separating hyper
plane with maximum margin between the two classes, which offer the good
generalization capacity for future data. Their storage and computation necessities
increment quickly with the quantity of preparing vectors. SVM utilizes statistical
increment quickly with the quantity of preparing vectors. SVM utilizes statistical
learning hypothesis to boost generalization property of produced classifier. The
“K Nearest Neighbor (KNN)” is a standout amongst the most generally utilized
classification algorithm in data mining, which is dependent of learning by
relationship, that is by contrasting a given test tuple and existing tuples which
resemble it. K-NN classification decides the decision boundary locally that was
produced from the need to perform segregate investigation when dependable
parametric assessments of probabilistic densities are obscure or hard to decide.

As in Figure 5 and Figure 6, a line is separated which called as hyper plane

which separates the two different classes clearly. In Figure 5 the distance
between the hyper plane and margin that is z1 and z2 should be equal. Hyper
plane can be found by kernel function. That is nothing but it finds the maximum
distance between the point’s plots in vector space. That exactly middle will be
the hyper plane. There is high probability that the points near to margin and
hyper plan to be get misclassified. In Figure 5 the point represented in hexagon
is to be classified. So, that point may get misclassified as it lies near to hyper
plane. In modified that is in the proposed algorithm misclassification can be
reduced by combining another technique to classify points near to the margin
and hyperplane. That is hyper plane not necessarily in exactly middle. Hyper
plane can be finding with logic for that we need get the ratio of count of data
points near to both margins. Depending upon ratio hyperplane distance from
both margins can be decided. Like in “k-nearest neighbor” algorithm we find the
class of new data point with comparing distance of new data point with other k
data points and from that k data points which class have more number of points
that class will be having new data point. So, the main disadvantage of k-NN is
we need to find distance of other points with respect to new data point. But in
proposed methodology we only need to find ratio only once and then finalize the
hyper plane.

Figure 5. Support vector machine

Figure 6. Proposed Support vector machine

New proposed methodology reduce misclassification without increasing the

New proposed methodology reduce misclassification without increasing the
actual execution time. The independent modules of the algorithm can be
executed parallel so that execution time can be reduced dramatically. This new
proposed methodology will analyze the existing data points which are near to the
boundary of the data points of that data class. Depending upon analysis the ratio
is to be considered. Data points of which class are more near to the boundary
area, then the hyper line shifted toward opposite so that misclassification can be
reduced. As per existing “Support Vector Machine” in (Figure 5) the new data
point represented by red colored hexagon belongs to different class of rectangles.
But same data point with “K-Nearest Neighbor” belongs to circular class. But
with new methodology this misclassification can be reduced as data points near
to boundary area are classified using two most popular classification algorithms
indirectly without increasing actual execution time.

SOLUTIONS AND RECOMMENDATIONS

Modified “support vector machine” reduces the miss classification without
increasing the actual execution time. As reference to the Figure 7 and Figure 8
mentioned below there is difference in the position of the hyper plane which is
separating the different class of data as represented with different colors. As
there are two colored data objects are there that is red and blue which is
separated by hyper plane. In Figure 7 that is existing SVM classification
algorithm’s scatterplot the margins are at equidistance from the margin. But
inverse to the existing classification algorithm as in Figure 8 the margins are not
at exactly same distance from the hyper plane. That distance will vary depending
upon the number data points nearer to the hyperplane. If as in Figure 8 data point
with border are considered for deciding the hyper plane and margin. Modified
“support vector machine” reduces the miss classification without increasing the
actual execution time. As reference to the Figure 7 and Figure 8 mentioned
below there is difference in the position of the hyper plane which is separating
the different class of data as represented with different colors. As there are two
colored data objects are there that is red and blue which is separated by hyper
plane. In Figure 7 that is existing SVM classification algorithm’s scatterplot the
margins are at equidistance from the margin. But inverse to the existing
classification algorithm as in Figure 8 the margins are not at exactly same
distance from the hyper plane. That distance will vary depending upon the
number data points nearer to the hyperplane. If as in Figure 8 data point with
border are considered for deciding the hyper plane and margin.
Figure 7. SVM Classifier Scatterplot

Figure 8. Modified SVM Scatterplot

FUTURE RESEARCH DIRECTIONS
This modified “Support vector machine” algorithm reduces the miss
classification but not remove completely. In future, more classification
algorithms can be added for more validation of data points. Like data points
which lies near to the hyper plan have more chances to get miss classified
similarly data points which lies near to boundaries are also can be easily get miss
classified.

CONCLUSION
This new proposed methodology reduce misclassification as it adjusts the hyper
plane depending upon data points around the border of the that class. This new
algorithm is nothing but the combination of two algorithms “Support Vector
Machine”, “K-Nearest Neighbour”. As there are high chances of
misclassification of data points which lies near to the boundary of that class. So,
to reduce the misclassification double verification of these sensitive area’s data
to reduce the misclassification double verification of these sensitive area’s data
point is to be done through this new algorithm. As there is parallel computation
of independent module is to be done, so execution time can be dramatically
decreased. This new proposed algorithm works without increasing actual
execution time but it reduces the misclassification.

REFERENCES

A Survey on Feature Selection Techniques and Classification

Algorithms for Efficient Text Classification. (2016). International
Journal of Science and Research.

Ali AlShaari, M. (2014). Text Documents Classification Using Word

Intersections. International Journal of Engineering and Technology ,
6(2), 119–122. doi:10.7763/IJET.2014.V6.678

Dayan, P. (n.d.). Unsupervised Learning. Appeared . In Wilson, R. A.,

& Keil, F. (Eds.), The MIT Encyclopedia Of The Cognitive Sciences .
MIT.

Dubey, V. (2016). Hybrid classification model of correlation-based

feature selection and support vector machine . IEEE.
doi:10.1109/ICCTAC.2016.7567338

Gayathri, K. (2013). Data pre-processing with the KNN for

classification using the SVM . IEEE.

Guang-chao, W. (2008). Support Vector Machine Classifier Based on

Fuzzy Partition and Neighborhood Pairs. Jisuanji Yingyong .

Gumus, F. (2014). Online Naive Bayes classification for network

intrusion detection . IEEE. doi:10.1109/ASONAM.2014.6921657

Kulesza, A. (2012). Determinantal Point Processes for Machine

Learning. Foundations And Trends® In . Machine Learning , 5(2-3),
123–286. doi:10.1561/2200000044

Kulesza, A. (2012). Determinantal Point Processes for Machine

Learning. Foundations And Trends® In . Machine Learning , 5(2-3),
123–286. doi:10.1561/2200000044

Laskov, P., & Lippmann, R. (2010). Machine learning in adversarial

environments. Machine Learning , 81(2), 115–119.
doi:10.1007/s10994-010-5207-6

Talwar, A., & Kumar, Y. (2013). Article. International Journal of

Engineering and Computer Science, 2(12).

Valsala, S., Ann George, J., & Parvathy, P. (2011). A Study of

Clustering and Classification Algorithms Used in Datamining.
International Journal of Computer Science and Network Security,
11(10).

Wang, L. (2012). Improved KNN classification algorithms research in

text categorization . IEEE. doi:10.1109/CECNet.2012.6201850

Wang, S. (2013). Support Vector Machines Classification for High-

Dimensional Dataset . IEEE.

Wu, G. (2008). Support vector machine classifier based on fuzzy

partition and neighborhood pairs. Jisuanji Yingyong , 28(1), 131–133.
doi:10.3724/SP.J.1087.2008.00131
 CHAPTER 4
An Analysis of Usage-Induced Big Data

Sameera K.
VIT University, India

P. Swarnalatha
VIT University, India

ABSTRACT
With the predominance of administration registering and distributed computing,
an ever-increasing number of administrations are developing on the internet,
producing tremendous volume of information. The mind-boggling
administration-created information turn out to be too extensive and complex to
be successfully prepared by customary methodologies. The most effective
method to store, oversee, and make values from the administration-situated
enormous information turn into a vital research issue. With the inexorably huge
measure of information, a solitary framework that gives normal usefulness to
overseeing and dissecting diverse sorts of administration-produced enormous
information is critically required. To address this test, this chapter gives a review
of administration-produced huge information and big data-as-a-service. Initially,
three sorts of administration-produced huge information are abused to upgrade
framework execution. At that point, big data-as-a-service, including big data
infrastructure-as-a-Service, big data platform-as-a-service, and big data analytics
software-as-a-service, is utilized to give regular huge information-related
administrations (e.g., getting to benefit-produced huge information and
information investigation results) to clients to improve effectiveness and lessen
cost.

INTRODUCTION
After entering the 21st century, the worldwide financial structure is exchanging
from “mechanical economy” to “administration economy”. As per the insights of
from “mechanical economy” to “administration economy”. As per the insights of
the World Bank, the yield of present day benefit industry takes more than 60
percent of the world yield, while the rate in created nations surpasses 70%. The
opposition in the region of current benefit industry is turning into a point of
convergence of the world's economy advancement. Benefit registering, which
gives adaptable registering designs to bolster present day benefit industry, has
developed as a promising exploration range. With the commonness of
distributed computing, increasingly present day administrations are conveyed in
cloud foundations to give rich functionalities. The quantity of administrations
and administration clients are expanding quickly. There has been gigantic blast
in information era by these administrations with the predominance of versatile
gadgets, client informal organizations, and substantial scale benefit situated
frameworks. The staggering administration produced information turn out to be
as well expansive and complex to be viably handled by conventional approaches.

In data innovation, huge information has developed as a broadly perceived

pattern, pulling in considerations from government, industry and the scholarly
world. Enormous information are high volume, high speed, and additionally high
assortment data resources that require new types of preparing to empower
upgraded basic leadership, understanding disclosure and process enhancement
(M. A. Beyer & D. Laney). Specified by the Compliance, Governance and
Oversight Council (CGOC, an association concentrated on Information
Governance), data volume copies each 18-24 months for most associations and
90% of the information on the planet has been made over the most recent two
years (D. Austin). In March 2012, the Obama organization reported the huge
information innovative work activity, which investigated how enormous
information could be utilized to address imperative issues confronting the
legislature. The driving IT organizations, for example, SAG, Oracle, IBM,
Microsoft, SAP and HP, have spent more than $15 billion on purchasing
programming firms represent considerable authority in information
administration and examination. This industry all alone is worth more than $100
billion and developing at right around 10% a year, which is generally twice as
quick as the product business in general (“A special report on managing
information: Data, data everywhere,”). Step by step instructions to proficiently
and viably make values from the huge information turn into an imperative
investigate issue.

The developing extensive scale benefit situated frameworks frequently include

countless with complex structures. The enormous information created from these
frameworks are ordinarily heterogeneous, of various information sorts, and very
powerful. Due to the quick increment of framework size and the related huge
powerful. Due to the quick increment of framework size and the related huge
volume of administration produced information, making an incentive in the
nearness of huge framework and information turns into an unavoidable test.
Cases of administration created enormous information incorporate follow logs,
Quality-of-Service (QoS) data, benefit summon relationship, and so forth. Like
different sorts of huge information, the usage-created huge information activities
traverse four remarkable measurements (“What is big data? ł bringing big data to
the enterprise,”): (1) volume: these days' huge scale frameworks are flooded with
constantly developing information, effortlessly gathering terabytes or even
petabytes of data; (2) speed: time-touchy procedures, such as bottleneck
identification and administration QoS forecast, could be accomplished as
information stream into the framework; (3) assortment: organized furthermore,
unstructured information are created in different information sorts, making it
conceivable to investigate new experiences while breaking down these
information together; and (4) veracity: identifying and remedying uproarious and
conflicting information are essential to lead trustable investigation. Building up
trust in enormous information displays an immense test as the assortment and
number of sources develops. These four one of a kind attributes of
administration created huge information give extraordinary test for information
administration and examination.

To satisfy the capability of administration produced huge information, creating

excellent innovations to viably handle huge amounts of information inside
adequate handling time is a basic undertaking. In addition, simple access of the
huge information and the enormous information examination results are
imperative. Huge Data-as-a-Service embodies different enormous information
stockpiling, administration, and investigation systems into administrations and
gives basic enormous information related administrations to clients by means of
programmable APIs, which extraordinarily upgrades proficiency, lessen cost and
empowers consistent reconciliation. To give enormous information foundation,
huge information stage, and huge information investigation programming
projects as administrations, there are a considerable measure of examine
examinations should be finished. This paper gives a diagram of administration
produced huge information and Big Data-as-a-Benefit. Initial, three sorts of
administration produced enormous information (benefit follow logs, benefit QoS
data, and administration relationship) are misused to improve framework
execution. At that point, Big Data-as-a-Service (BDaaS) is examined to give
inviting APIs to clients to get to the administration created enormous
information and information investigation comes about.
Whatever is left of this paper is sorted out as takes after: Section 2 gives a
diagram of this paper; Section 3 misuses benefit produced enormous
information; Section 4 explores Big Data-as-a-Service; Section 5 investigates
business parts of usage-created huge information and Big Data-as-a-Service and
Section 6 closes the paper.

ANALYSIS
Figure 1 gives an outline of the administration created enormous information
and Big Data as-a-Service. As appeared in the figure, on the one hand, we will
present some run of the mill applications which misuse three sorts of
administration produced huge information separately for framework execution
upgrade. In the first place, log representation what's more, execution program
conclusion are examined by means of mining administration ask for follow logs.
Second, QoS-mindful blame resilience and administration QoS forecast are
concentrated in light of the benefit QoS data. At long last, noteworthy
administration recognizable proof also, benefit relocation are accomplished by
examining administration relationship. These solid applications will reveal some
insight on the issue of enormous information investigation by mining the usage-
created huge information.

Then again, to give easy to understand access to the administration produced

huge information and different huge information logical comes about, Big Data-
as-a-Service will likewise be presented as a fundamental structure to store,
oversee and make an incentive from the huge information. Figure 1 shows the
structure of Big Data-as-a-Service, which includes three layers, i.e., Big Data
Infrastructure-as-a-Benefit, Big Data Platform-as-a-Service, and Big Data
Analytics Programming as-a-Service. By means of standard and programmable
APIs, Big Data-as-a-Service empowers dynamic joining of diverse huge
information and combination of various huge information examination ways to
deal with make an incentive from the administration created huge information.

Figure 1. Analysis of Usage-induced Big Data and Big Data-as-a-Service

USAGE-INDUCED BIG DATA
These days, there are a wide range of online administrations gave on the Internet
and every day utilized by a huge number of clients. Each time when you play out
an inquiry, send an Email, post a micro blog or, on the other hand shop on online
business Websites, you are creating a follow of information to the
administrations.

As appeared in Figure 2, as the quantity of administrations and clients scales up,

the administration produced information (counting administration logs, benefit
QoS data, and administration relationship) are expanding, prompting the huge
information marvel. With the expanding volume of administration produced
huge information, how to make values from the information turns into a critical
research issue. The accompanying sub-areas will depict in detail how the benefit
produced information can be prepared and broke down to upgrade framework
execution.

Figure 2. Usage-induced Big Data

A. Usage-Indicated Registers

With the promotion of expansive scale benefit situated frameworks, also, the
number expanding of administration clients (e.g., PCs, cell phones, and so on.), a
gigantic volume of follow logs are produced by the administration arranged
frameworks every day. There are billions of every day logs, log documents, and
organized/unstructured information from a wide assortment of administration
frameworks. For instance, an Email benefit given by Alibaba (one of the greatest
web based business organization on the planet) would deliver around 30-50
gigabytes (around 120-200 million lines) of following logs every hour (H. Mi,
H.Wang, Y. Zhou, M. R. Lyu, & H. Cai). These logs can be used both in the
improvement stages and in ordinary operations for comprehension and
investigating the conduct of the perplexing framework.

As the scale and multifaceted nature of disseminated frameworks quickly

increment, extensive scale disseminated frameworks like distributed computing
frameworks ordinarily include countless between benefit segments, now and
again crosswise over many machines, which makes it extremely hard to
physically analyze the execution issues. By the method for mining these usage-
created follow logs, important data can be acquired to help benefit originators
and engineers comprehend and move forward the nature of frameworks. For
instance, execution bottleneck confinement can be accomplished by breaking
down the follow logs. In any case, there are as yet many difficulties to be tended
to. On one hand, the enormous volume of administration produced follow logs
make execution finding work concentrated; then again, the interest for
continuous framework analysis is continually expanding.

This area talks about how to research the follow logs to discover the esteem
This area talks about how to research the follow logs to discover the esteem
covered up in it, including follow log perception furthermore, execution issue
conclusion.

1. Induced Register Envision: Disseminated frameworks are constantly

developing in scale and administration segment collaborations. It is getting
to be noticeably troublesome for the framework planners and chairmen to
comprehend the attributes of framework execution. Ask for following
methodologies, which are broadly embraced by organizations, for example,
Google (B. H. Sigelman, L. A. Barroso, M. Burrows, P. Stephenson, M.
Plakal, D. Beaver, S. Jaspan, & C. Shanbhag), Microsoft (S. Han, Y. Dang,
S. Ge, D. Zhang, & T. Xie) and eBay (M. Y. Chen, A. Accardi, E. Kiciman,
J. Lloyd, D. Patterson, A. Fox, & E. Brewer), record the execution data of
solicitations (e.g., entering also, leaving time when individual solicitations
experience benefit parts). These follow logs contain a great deal of
shrouded fortunes for framework architects and overseers. Log
representation gives apparatuses to digest representation of log records (or
results of log inquiries) in a way that can be best comprehended by clients
(C. Lim, N. Singh, & S. Yajnik). With the expanding volume of follow
logs, effective log representation of turn into a test examine issue.

To address this issue, various methodologies have been proposed. Stardust (E.
Thereska, B. Salmon, J. Strunk, M. Wachs, M. Abd-El-Malek, J. Lopez, & G. R.
Ganger) utilizes social databases as store which endures poor inquiry
effectiveness in the earth of huge information volumes. Ptracer (H. Mi, H.
Wang, H. Cai, Y. Zhou, M. R. Lyu, & Z. Chen) is an online execution profiling
device to pictures multi-dimensional measurable data to help chairmen
comprehend the framework execution practices top to bottom. DTrace (B. M.
Cantrill, M. W. Shapiro, & A. H. Leventhal) and gprof (S. L. Graham, P. B.
Kessler, & M. K. Mckusick) picture the execution of frameworks as call charts
to connote where demands invest energy.

In spite of the fact that various past research examinations have been directed at
administration log representation, this exploration issue turns out to be all the
more difficult in the situation of enormous information, created by the quick
increment of log documents, the unstructured log information, and the
prerequisite of continuous inquiry and show. More research examinations are
expected to empower constant handling and representation of the enormous
volume of follow logs.
 2. Detection of Difficulties in Functioning: In today's disseminated
frameworks, particularly the cloud frameworks, an administration demand
will experience diverse hosts, conjuring various programming modules. At
the point when the administration can't fulfill the guaranteed benefit level
understanding (SLA) to clients, it is basic to distinguish which module
(e.g., a summoned technique) is the underlying driver of the execution issue
in an opportune way. Follow logs give important data to discover the reason
for execution issues. Step by step instructions to misuse the gigantic follow
logs successfully what's more, effectively to help creator comprehend
framework execution what's more, find execution practices turns into a
pressing and testing research issue.

In late writing, an expansive assortment of research work has been researched to

address this issue. For instance, Magpie (P. Barham, A. Donnelly, R. Isaacs, &
R. Mortier), Pip (P. Reynolds, C. Killian, J. L. Wiener, J. C. Mogul, M. A. Shah,
& A. Vahdat) and Iron model (E. Thereska & G. R. Ganger) are application-
particular approaches, which require space learning to build the execution
conclusion models. Pinpoint (M. Chen, E. Kiciman, E. Fratkin, A. Fox, & E.
Brewer) utilizes bunching calculation to gathering disappointment and
achievement logs. Spruce (B. H. Sigelman, L. A. Barroso, M. Burrows, P.
Stephenson, M. Plakal, D. Beaver, S. Jaspan, & C. Shanbhag) utilizes Bigtable
to deal with the extensive volume of follow logs. Information mining
innovations, for example, chief part examination (PCA) and powerful chief part
investigation are additionally utilized for distinguishing execution bottlenecks
through demonstrating and mining the follow logs (H. Mi, H.Wang, Y. Zhou, M.
R. Lyu, & H. Cai), (L. Shao, J. Zhang, Y. Wei, J. Zhao, B. Xie, & H. Mei,
2007).

In any case, the vast volume and high speed of usage-created follow logs make it
extremely hard to perform real-time analysis. The vast majority of the past
arrangements experience the ill effects of low proficiency in dealing with
expansive volume of information. More effective capacity, administration, and
investigation approaches for usage-created follow logs are required.

B. Usage QoS Data

The present expansive scale dispersed stages (e.g., different cloud stages) give
various administrations to heterogeneous also, expanded clients. Extensive
volume of QoS information of these administrations are recorded, in both server-
side and client side. Since distinctive clients may watch very unique QoS
side and client side. Since distinctive clients may watch very unique QoS
execution (e.g., reaction time) on a similar administration, the volume of user-
side QoS information is considerably bigger than that of server-side QoS
information. In addition, QoS estimations of administration segments are
evolving progressively every once in a while, bringing about touchy increment
of client side administration QoS data.

Profitable data can be acquired through researching these client side

administration QoS data keeping in mind the end goal to upgrade framework
execution, for instance, to accomplish versatile blame resilience and to make
customized QoS expectation for clients.

1. Flaw Deviation: The administration figuring condition is exceedingly

unique and heterogeneous, where unique administrations might be
incapacitated, new administrations might be included, and QoS of the
administrations may change every now and then. Building solid benefit
situated frameworks is substantially more difficult in this exceedingly
powerful condition contrasted and the conventional remain solitary
programming frameworks. Programming adaptation to internal failure (M.
R. Lyu) is a vital way to deal with construct dependable frameworks by
means of utilizing practically comparable parts to endure deficiencies. On
the Web, the practically identical Web administrations gave by various
associations can be utilized to manufacture fault-tolerant benefit arranged
frameworks. When planning adaptation to internal failure techniques,
benefit QoS data can be considered to improve the execution.

The colossal number of administrations in the expansive scale circulated stages

is observed ceaselessly at runtime. Extensive volume of QoS information (e.g.,
reaction time, accessibility, throughput, and so forth.), is recorded. In addition,
subsequent to leading administration summon, the clients likewise record the
QoS of the summoned administrations. How to proficiently and viably handle
these extensive volume of benefit QoS information to configuration adaptation
to internal failure systems which can adjust to the dynamic condition for ideal
execution is a testing research issue.

In our past work (Z. Zheng & M. R. Lyu), a preparatory middleware has been
intended for blame tolerant Web administrations. Be that as it may, this
middleware did not give a customized adaptation to internal failure system for
various clients. In the dynamic Internet condition, server-side adaptation to non-
critical failure is insufficient since the correspondence associations can bomb
critical failure is insufficient since the correspondence associations can bomb
effortlessly. Customized user-side adaptation to non-critical failure should be
considered. In addition, to speed up the investigation and calculation of the huge
volume of benefit QoS data, web based learning calculations (H. Yang, Z. Xu, I.
King, & M. Lyu) will should be explored for incremental refresh of the blame
resistance procedure when new QoS values end up plainly accessible.

2. QoS Prognosis: Web benefit QoS forecast goes for giving customized
QoS esteem forecast to administration clients, by utilizing the verifiable
QoS estimations of various clients. Web benefit QoS forecast more often
than excludes a client benefit framework, where every passage in the
network speaks to the estimation of a specific QoS property (e.g., reaction
time) of a Web benefit seen by an administration client. The client benefit
lattice is normally exceptionally inadequate with many missing sections,
since an administration client normally just conjured few Web benefits in
the past. The issue is the means by which to precisely foresee the missing
QoS values in the client benefit lattice by utilizing the accessible QoS
values. Subsequent to foreseeing the missing Web benefit QoS values in the
client benefit framework, each administration client can have a QoS
assessment on every one of the administrations, even on the unused
administrations. Thus, ideal administration can be chosen for clients to
accomplish great execution.

In administration processing, Web benefit QoS expectation has pulled in a

considerable measure of consideration as of late. Various QoS expectation
approaches have been proposed to address this paper, counting client based QoS
forecast approach (L. Shao, J. Zhang, Y. Wei, J. Zhao, B. Xie, & H. Mei), mix of
client based and thing based methodologies (Z. Zheng, H. Ma, M. R. Lyu, & I.
King), ranking-oriented approach, (Z. Zheng, H. Ma, & M. R. Lyu,) and I. King
bunching based approach (X. Chen, Z. Zheng, X. Liu, Z. Huang, & H. Sun), and
so on. Some current work (M. Tang, Y. Jiang, J. Liu, & X. F. Liu), (W. Lo, J.
Yin, S. Deng, Y. Li, & Z. Wu) additionally considers the areas of Web
administrations and administration clients to improve the unique situation data of
the administration condition and accordingly make strides the expectation
quality. Nonetheless, with the quick expanding of Web administrations and
clients, the measure of client administration network is getting to be plainly
bigger and bigger, which makes it not productive for continuous expectation.
Since the Internet condition is exceptionally dynamic, administrations may
include or drop at whatever time, so improving the vigor of QoS expectation
approaches (e.g., settling the frosty begin issue) is exceptionally basic. What's
more, estimations of some client side QoS properties (e.g., reaction time) are
more, estimations of some client side QoS properties (e.g., reaction time) are
changing after some time, making the lattice turn into a three-dimensional user
service-time network. Along these lines, how to effectively prepare the huge
volume of accessible administration QoS information and precisely anticipate
the missing QoS values in the colossal client benefit time lattice turns into an
exceptionally difficult research issue.

C. Usage Alliance

These days, substantial scale dispersed frameworks regularly include a

substantial number of administration segments. These administration segments
are commonly conveyed in disseminated PC hubs (i.e., physical machines or
virtual machines) and have complex conjuring connections. For instance, to
produce the dynamic Web content for a page in one of the internet business
destinations in Amazon, each demand regularly requires the page rendering
segments to build its reaction by sending solicitations to more than 150
administrations (G. DeCandia, D. Hastorun, M. Jampani, G. Kakulapati, A.
Lakshman, A. Pilchin, S. Sivasubramanian, P. Vosshall, & W. Vogels). These
segments are related between each other. The conjuring relationship among
administration parts can be displayed as a weighted coordinated diagram, where
a hub in the chart speaks to an administration segment also, a guided edge
starting with one hub then onto the next speaks to a part summon relationship.
The weight at each edge might be communicated as the cost or the recurrence of
the conjuring. This administration summon chart can be refreshed progressively
at runtime, brought on by reasons, for example, benefit include/drop, benefit
relocation, stack adjust, and so on.

By misusing the administration summon chart, important data can be gotten to

distinguish noteworthy administration parts what's more, to empower better
powerful administration relocation.

1. Recognition of Eloquent Usage Constituent: Unwavering quality of

various administration segments may force distinctive impacts on the
administration situated framework. By examining the benefit summon
chart, the critical administration segments (e.g., center parts or feeble
segments) can be recognized. This can significantly help us see how to
enhance the structure of a framework and how to enhance the unwavering
quality of the framework. For instance, extra adaptation to non-critical
failure techniques can be intended for these huge administration parts to
 accomplish higher unwavering quality. Nonetheless, because of the way of
dynamic creation of administration segments, the administration conjuring
chart can be ceaselessly refreshed at runtime. Also, the pattern towards
huge scale frameworks make the administration summon chart very
substantial and complex. Stochastic positioning procedures can be utilized
to distinguish the critical administration part in the chart for an appropriated
framework.

In our preparatory examination, we considered a segment huge on the off chance

that it is summoned by numerous other imperative segments as often as possible,
and proposed an irregular walk based way to deal with recognize critical
segments in a cloud framework. Plus, Liu et al. (A. Liu, Q. Li, L. Huang, & S.
Wen) propose to misuse the administration relationship to help the notoriety
calculation of administrations, which additionally enhance the strength of the
framework. Be that as it may, there are still a ton of research issues to be tended
to, for instance: (1) demonstrating the connection between various benefit parts
(e.g., segments an and b conjure the same part c, then there is understood
connection between part a and b; (2) demonstrating the effect of segment on the
entire framework, which can help us enhance the heartiness of entire framework;
(3) planning more proficient and successful ways to deal with fabricate and
dissect the administration conjuring chart what's more, distinguish huge
administration parts.

2. Usage Expatriation: Dispersed frameworks ordinarily incorporate

various administration segments, which should be sent to circulated hubs
(i.e., physical machines or virtual machines). Since the administration
condition is very powerful, after the underlying sending, it is basic to
enhance the benefit sending methodology among hopeful hubs
intermittently to accomplish ideal general framework execution while
limiting the operational cost. Accordingly, dynamic administration
movement is in need by moving the administration from one physical
machine to another at runtime. It is a typical practice in numerous business
cloud stages.

By demonstrating and misusing the administration summon relationship what's

more, past administration use encounters, a legitimate movement of the
administrations can enhance the experience for existing clients. In our past work
(Y. Kang, Z. Zheng, & M. Lyu), a preparatory model has been defined in light of
whole number programming to make an ideal redeployment of administrations.
In any case, this number programming based model experiences the scaling issue
In any case, this number programming based model experiences the scaling issue
when confronting countless and competitor hubs. This model is additionally
enhanced in (J. Zhu, Z. Zheng, Y. Zhou, & M. R. Lyu) by taking administration
relationship into account, which proposes to utilize a hereditary calculation to
settle the demonstrate productively. With the expansion of cloud organization,
some other work (e.g., (Q. Zhang, Q. Zhu, M. F. Zhani, & R. Boutaba; M.
Steiner, B. G. Gaglianello, V. K. Gurbani, V. Hilt, W. D. Roome, M. Scharf, &
T. Voith, M. Alicherry & T. V. Lakshman) additionally considers the dynamic
administration arrangement and relocation in geologically appropriated mists to
get ideal the administration execution.

To adapt to the developing size of the administration movement issue, more

productive methodologies are required. For instance, (1) notwithstanding
considering the conjuring relationship among administrations, areas of
administration clients can be considered to additionally enhance the movement
execution; (2) utilizing QoS expectation procedures, arrange latencies among
administration segments and amongst clients and administrations can be
anticipated to accomplish better administration movement execution; (3) to
speed up the calculation of ideal administration relocation procedure, of number
programming, different machine learning calculations, for example, web based
learning, and k-middle enhancement show, hereditary calculation, and so on.,
can be examined to empower proficient benefit movement.

CONCLUSION AND FUTURE WORK

This paper gives a diagram of administration produced huge information and Big
Data-as-a-Service. Three sorts of usage-created huge information are misused to
upgrade nature of service-oriented frameworks. To give basic usefulness of
enormous information administration and examination, Big Data-as-a-Service is
explored to give APIs to clients to get to the administration created enormous
information and the huge information examination comes about.

Later on, next to the administration follow logs, QoS data what's more,
administration relationship, more sorts of administration created huge
information will be researched. More complete investigations of different
administration produced huge information investigation methodologies will be
led. Point by point innovation guide will be given what's more, security issues
past the extent of this paper will likewise be explored.
REFERENCES

Alicherry, M., & Lakshman, T. V. (2012). Network aware resource

allocation in distributed clouds . Proceedings - IEEE INFOCOM , 12,
963–971.

Austin. (2012). eDiscovery Trends: CGOCs Information Lifecycle

Governance Leader Reference Guide. Retrieved from
http://www.ediscoverydaily.com

BarhamP.DonnellyA.IsaacsR.MortierR. (2004). Using magpie for

request extraction and workload modelling. Proceedings of the 6th
conference on Symposium on Opearting Systems Design &
Implementation (OSDI’04), 18–18.

Beyer & Laney. (2012). The importance of ‘big data’: A definition .

Gartner.

CantrillB. M.ShapiroM. W.LeventhalA. H. (2004). Dynamic

instrumentation of production systems. USENIX Annual Technical
Conference, 15–28.

ChenM.KicimanE.FratkinE.FoxA.BrewerE. (n.d.). Pinpoint: problem

determination in large, dynamic internet services. Proceedings of the
International Conference on Dependable Systems and Networks
(DSN’02), 595–604. 10.1109/DSN.2002.1029005

ChenM. Y.AccardiA.KicimanE.LloydJ.PattersonD.FoxA.BrewerE.
(2004). Path-based faliure and evolution management. In Proceedings
of the 1st conference on Symposium on Networked Systems Design
and Implementation. USENIX Association.

Chen, X., Zheng, Z., Liu, X., Huang, Z., & Sun, H. (2011).
Personalized QoS-aware Web service recommendation and
visualization. IEEE Transactions on Services Computing .

DeCandia, Hastorun, Jampani, Kakulapati, Lakshman, Pilchin, …

Vogels. (2007). Dynamo: amazon’s highly available key-value store.
Proc. 21st ACM Symposium on Operating Systems Principles
(SOSP’07), 205–220.

Graham, S. L., Kessler, P. B., & Mckusick, M. K. (1982). Gprof: A

call graph execution profiler . ACM SIGPLAN Notices , 17(6), 120–
126. doi:10.1145/872726.806987

HanS.DangY.GeS.ZhangD.XieT. (2012). Performance debugging in

the large via mining millions of stack traces. Proc. 34th Int’l Conf. on
Software Engineering (ICSE’12), 145–155.
10.1109/ICSE.2012.6227198

IBM. (2013). What is big data? ł bringing big data to the enterprise.
Retrieved from http://www-01.ibm.com/software/data/bigdata

KangY.ZhengZ.LyuM. (2012). A latency-aware co-deployment

mechanism for cloud-based services. Proceedings of the IEEE 5th
International Conference on Cloud Computing (CLOUD’12), 630–
637. 10.1109/CLOUD.2012.90

LimF.SinghN.YajnikS. (2008). A log mining approach to failure

analysis of enterprise telephony systems. Proceedings of the IEEE
International Conference on Dependable Systems and Networks
(DSN’08), 398–403.

Liu, Li, Huang, & Wen. (2012). Shapley value based impression
propagation for reputation management in web service composition.
Pro. IEEE 19th Int’l Conf’ on Web Services (ICWS’12), 58–65.

Lo, W., Yin, J., Deng, S., Li, Y., & Wu, Z. (2012). Collaborative web
service qos prediction with location-based regularization. Pro. IEEE
19th Int’l Conf’ on Web Services (ICWS’12), 464–471.
10.1109/ICWS.2012.49

Lyu, M. R. (1995). Software Fault Tolerance. In Trends in Software .

Wiley.

MiH.WangH.CaiH.ZhouY.LyuM. R.ChenZ. (2012). Ptracer: Path-

based performance profiling in cloud computing systems. In
Proceedings of the 36th IEEE Annual Computer Software and
Applications Conference (COMPSAC’12). IEEE.

Mi, Wang, & Zhou, Lyu, & Cai. (2013). Towards fine-grained,
unsupervised, scalable performance diagnosis for production cloud
computing systems. IEEE Transactions on Parallel and Distributed
Systems .

ReynoldsP.KillianC.WienerJ. L.MogulJ. C.ShahM. A.VahdatA.

(2006). Pip: detecting the unexpected in distributed systems.
Proceedings of the 3rd conference on Networked Systems Design &
Implementation (NSDI’06), 9–9.

ShaoL.ZhangJ.WeiY.ZhaoJ.XieB.MeiH. (2007). Personalized QoS

prediction for Web services via collaborative filtering. Proc. 5th Int’l
Conf. Web Services (ICWS’07), 439–446. 10.1109/ICWS.2007.140

Sigelman, Barroso, Burrows, Stephenson, Plakal, Beaver, …

Shanbhag. (2010). Dapper, a large-scale distributed systems tracing
infrastructure. Google, Inc.

SteinerM.GaglianelloB. G.GurbaniV. K.HiltV.RoomeW.

D.ScharfM.VoithT. (2012). Network-aware service placement in a
distributed cloud environment. Proc. ACM SIGCOMM’12, 73–74.
10.1145/2342356.2342366
Tang, M., Jiang, Y., Liu, J., & Liu, X. F. (2012). Location-aware
collaborative filtering for qos-based service recommendation. Pro.
IEEE 19th Int’l Conf’ on Web Services (ICWS’12), 202–209.

The Economist. (2010). A special report on managing information:

Data, data everywhere. The Economist.

Thereska, G., Salmon, B., Strunk, J., Wachs, M., Abd-El-Malek, M.,
Lopez, J., & Ganger, G. R. (2006). Stardust: tracking activity in a
distributed storage system. ACM SIGMETRICS Performance
Evaluation Review, 34(1), 3–14. doi:10.1145/1140277.1140280

ThereskaH.GangerG. R. (2008). Ironmodel: robust performance

models in the wild. Proceedings of the International Conference on
Measurement and Modeling of Computer Systems (SIGMETRICS
’08), 253–264. 10.1145/1375457.1375486

XuW.HuangL.FoxA.PattersonD.JordanM. I. (2009). Detecting large-

scale system problems by mining console logs. Proceedings of the
ACM 22nd Simposium on Operating Systems Principles (SOSP’09),
117–132. 10.1145/1629575.1629587

YangH.XuZ.KingI.LyuM. (2010). Online learning for group lasso.

International Conference on Machine Learning (ICML’10).

ZhangQ.ZhuQ.ZhaniM. F.BoutabaR. (2012). Dynamic service

placement in geographically distributed clouds. Proc. IEEE 32nd Int’l
Conf. on Distributed Computing Systems (ICDCS’12), 526–535.
10.1109/ICDCS.2012.74

ZhengZ.LyuM. R. (2009). A QoS-aware fault tolerant middleware for

dependable service composition. Proc. 39th Int’l Conf. Dependable
Systems and Networks (DSN’09), 239–248.
10.1109/DSN.2009.5270332
Zheng, Z., Ma, H., Lyu, M. R., & King, I. (2011). QoS-aware Web
service recommendation by collaborative filtering . IEEE Transactions
on Services Computing , 4(2), 140–152. doi:10.1109/TSC.2010.52

ZhengZ.ZhangY.LyuM. R. (2010). CloudRank: A QoS-driven

component ranking framework for cloud computing. Proc. Int’l Symp.
Reliable Distributed Systems (SRDS’10), 184–193.
10.1109/SRDS.2010.29

ZhengZ.ZhuJ.LyuM. R. (2013). Usage-created Big Data and Big

Data-as-a-Service: An Overview. IEEE International Congress on Big
Data.

ZhuJ.ZhengZ.ZhouY.LyuM. R. (2013). Scaling service-oriented

applications into geo-distributed clouds. Pro. IEEE Int’l Workshop on
Internet-based Virtual Computing Environment (iVCE’13).
 CHAPTER 5
Glorified Secure Search Schema Over Encrypted Secure Cloud
Storage With a Hierarchical Clustering Computation

Shweta Annasaheb Shinde

VIT University, India

Prabu Sevugan
VIT University, India

ABSTRACT
This chapter improves the SE scheme to grasp these contest difficulties. In the
development, prototypical, hierarchical clustering technique is intended to lead
additional search semantics with a supplementary feature of making the scheme
to deal with the claim for reckless cipher text search in big-scale surroundings,
such situations where there is a huge amount of data. Least relevance of
threshold is considered for clustering the cloud document with hierarchical
approach, and it divides the clusters into sub-clusters until the last cluster is
reached. This method may affect the linear computational complexity versus the
exponential growth of group of documents. To authenticate the validity for
search, minimum hash sub tree is also implemented. This chapter focuses on
fetching of cloud data of a subcontracted encrypted information deprived of loss
of idea and of security and privacy by transmission attribute key to the
information. In the next level, the typical is improved with a multilevel
conviction privacy preserving scheme.

INTRODUCTION
Individuals are profited with cloud computing as cloud computing reduces it
work and make computing and storage simplified. (Liang, Cai, Huang, Shen &
Peng, 2012), (Mahmoud & Shen, 2012), (Shen, Liang, Shen, Lin & Lou, 2012).
Data can be stored remotely in the cloud server as data outsourcing and accessed
publicly. This embodies a mountable, constant and low-cost method for public
publicly. This embodies a mountable, constant and low-cost method for public
access of data as per the high productivity and mount ability of cloud servers,
and so it is favored.

Sensitive privacy information is of concern. Data should be encrypted before

sending to the cloud servers (Jung, Mao, Li, Tang, Gong & Zhang, 2013),
(Yang, Li, Liu & M, 2014). The data encryption comes with it the difficulty of
searching the data on the cloud servers. (Cao, Wang, Li,Ren & Lou, 2014)
Encryption comes with it many of other security apprehensions. Secure Sockets
Layer is used by Google search to encrypt the connection be the authors the
google server and search user.

Nevertheless, if the user clicks from the authors site of search result, to another
the authors site will identify the search terms the user has used.

On dealing with the above matters, the searchable form of encryption (e.g.,
(Song,Wagner & Perrig, 2000), (Li,Xu,Kang,Yow & Xu, 2014),
(Li,Lui,Dai,Luan & Shen, 2014)) has been established as a basic method to
allow searching over encrypted data of cloud, which profits the procedures. At
first the owner of data will produce quite a few keywords rendering to the
outsourced data. Cloud server will be used to store this encrypted keywords.
When the outsourced data needs to be accessed, it can choice approximately
appropriate keywords and direct the cipher text of the designated keywords to
the cloud server. The cloud server then usages the cipher text to contest the
outsourced keywords which are encrypted, and finally will yields the matching
consequences to the user who search. To attain the like search effectiveness and
accuracy over data which is encrypted as like plaintext search of keyword, a
widespread form of research has been advanced in literature. Wang et al.(2014)
recommended a ranked keyword search system which deliberates the scores of
relevance’s of keywords. Inappropriately, because of using order-preserving
encryption (OPE)(Boldyreva,Chenette, Lee & Oneill, 2009) to attain the
property of ranking, the planned arrangement cannot attain unlikability of
trapdoor.

Later, Sun et al.(Sun,Wang,Cao,Li,Lou,Hou & Li, 2013)suggested a

multikeyword text search arrangement which deliberates the scores of
relevance’s of the keywords and exploits a multidimensional tree method to
realize the authors organized query of search. (J. Yu, P. Lu, Y. Zhu, G. Xue, &
M. Li, 2013)suggested a multikeyword top-k retrieval organization which
practices fully homomorphic encryption to encrypt the index/trapdoor and
assurances high security. Cao et al. (2014) suggested a multikeyword ranked
search (MRSE), which put on machine of coordinate as the matching of keyword
rule, i.e., it will return the data with the maximum matching of keywords. Even
though many of the functionalities of search have been advanced in former
literature on the way to exact and the authors organized searchable encryption, it
is still problematic for searchable encryption to attain the similar user
involvement as that of the plaintext search, like Google search. This mostly
attributes to subsequent two issues. At first, query with user favorites is popular
in the search of plaintext (Liang, Cai,Huang,Shen, & Peng, 2012), (Mahmoud &
Shen, 2012). It allows tailored search and can more precisely represent
requirements of users, but has not been methodically studied and maintained in
the encrypted domain of data. At second, to further improve the user’s
experience on searching, a significant and vital function is to allow the
multikeyword search with the comprehensive logic operations, i.e., the “AND”,
“OR” and “NO” operations of keywords. This is vital for search users to trim the
space of searching and rapidly classify the anticipated data.

Cao et al. advise the coordinate matching search scheme (MRSE) which can be
the authors as a searchable encryption system with “OR” operation (Shen,Liang,
Shen,Lin, & Luo, 2014) recommended a conjunctive keyword search scheme
which can be observed as a searchable encryption scheme with “AND”
operation with the refunded documents matching all keywords. Though, most
current suggestions can only allow search with single logic operation, somewhat
than the mixture of numerous logic operations on keywords, which encourages
the work.

Here, the authors discourse above two issues by emerging two Fine-grained
MultiKeyword Search (FMS) arrangements over encrypted data of cloud. Our
unique donations can be abridged in three characteristics as tracks:

• The authors familiarize the relevance scores of relevance’s and the

preference factors of keywords for searchable encryption. The scores of
relevance’s of keywords can allow more detailed refunded consequences,
and the factors of preferences of keywords signify the standing of keywords
in the search for keyword set quantified by search of users and consistently
the authors search at personalized level to cater to precise preferences of
users. Thus, additional advances the search of functionalities and
experience of user.
• The authors understand the “AND”, “OR” and “NO” processes in the
multikeyword search for searchable encryption. Associated with
 multikeyword search for searchable encryption. Associated with
arrangements in the proposed arrangement can accomplish more all-
inclusive functionality and the authors query complexity of query.
• The authors employment the classified sub-dictionaries technique to
improve the effectiveness of the above two arrangements. Extensive
experimentations establish that the enhanced arrangements can attain better
effectiveness in terms of building of index, trapdoor generating and query
in the judgement with arrangements in.

Hardware restriction of mobile devices is overcome by Mobile cloud computing

(Dinh, Lee, Niyato, & Wang, 2013; Li, Dai, Tian, & Yang, 2009) by discovering
the accessible and virtualized storage of cloud and computing resources, and
consequently can deliver much more significant and mountable services of
mobile to user. In the technology of mobile cloud computing, mobile operators
characteristically are outsourcing their information to cloud servers which are
external, e.g., iCloud, to adore a steady, low-cost and climbable way for storage
of data and access. Though, as outsourced data has sensitive private information,
such as personal photos, emails., which would lead to severe confidentiality and
privacy violations (W. Sun, 2013), if without efficient protections. It is therefore
essential to encrypt the sensitive data before outsourcing them to the cloud. The
data encryption, the authors would result in salient problems when other users
need to access interested data with search, due to the problems of search over
encrypted data. This fundamental issue in mobile cloud computing consequently
inspires an extensive body of investigation in the recent years on the
examination of searchable encryption performance to attain the authors ll-
organized thorough over outsourced encrypted data (Wang, Lou, & Hou, 2014;
Yang, Liu, & Yang, 2014)

An assortment of research the whole thing have freshly been established about
multikeyword search over the data which is encrypted. Cash et al.
(Jarecki,Jutla,Krawczyk,Ro3u, & Steiner, 2013) offer a symmetric searchable
encryption organization which attains high effectiveness for big databases with
uncertain on security guarantees. Cao et al. (Cao, Wang, Li, Ren, & Lou,
2014)suggest a multikeyword search structure supportive consequence ranking
by approving k-nearest neighbors (kNN)technique (Wong, Cheung,Kao, &
Mamoulis,2009). Naveed et al. (2014) proposition an active searchable
encryption system complete blind storage to obscure admittance pattern of the
user for search. In demand to encounter the practical search necessities, search
concluded data which is encrypted should provision the subsequent three
functions. First, the encryption systems which are searchable should provision
multikeyword search, and deliver the same experience for user as thorough in
search for Google with different keywords; search for single-keyword is far from
acceptable by only recurring very incomplete and imprecise results for results.
Second, to rapidly classify most applicable results, the user for search would
characteristically favor cloud servers to category the refunded search
consequences in a relevance-based command (Pang, Shen, & Krishnan, 2010)
ranked by the order of relevance of the request for search documents. In
accumulation, display the search based on rank to users can also eradicate the
needless traffic of network by only distributing back the utmost results which are
relevant from to search users from cloud. Third, as for the effectiveness of
search, then the quantity of the documents which are imperfect in a database
could be tremendously large, encryption which is searchable constructions
should be organized in the authors-mannered to quickly response to the
requirements for search with interruptions and they are smallest. In modification
to the proposed prosperities, frequently of the usual proposals, the authors, nose
dive to proposal satisfactory intuitions near the construction of full performed
encryption which is searchable. As an application near the subject, the authors
proposition and the authors-organized multikeyword ranked search (EMRS)
preparation over encrypted cloud data for mobile through blind storage. Our
important charities can be abridged as surveys:

• The authors explain a relevance for score in encryption which is

searchable to accomplish multikeyword ranked search finished the cloud
data for mobile which is encrypted. In gathering to that, the authors
proposition and the authors organized index to advance the efficiency for
search.
• By adapting the blind storage scheme in the Meurthe authors resolve the
trapdoor unlikability problematic and obscure admittance pattern of the user
for search from the cloud server.
• The authors give systematic analysis of security to prove that the EMRS
can spread a high security level counting documents for confidentiality and
index, privacy with trapdoor, trapdoor unlikability, and covering admission
pattern of the search user. Moreover, the authors implement extensive
experiments, which show that the EMRS can achieve enhanced efficiency
in the terms of functionality and search effectiveness compared with
prevailing proposals.

Cloud computing is known as a substitute to outdated information technology

and has been progressively familiar as the greatest significant revolving point in
and has been progressively familiar as the greatest significant revolving point in
the expansion of information technology owing to its inherent sharing of
resource and low maintenance characters. Cloud computing is a computing
model, where the information which is shared, resources and software are
supplied to devices and computers based on requirement. This allows the end
user to admittance the resources for cloud computing anytime from required
platform such as mobiles, or desktops which is the mobile computing platform.

Clouds are huge pools of easily practical and available resources which are
virtualized. The data and the applications which are software essential by the
workers are not stored on their self-computers; in its place, they are the authors
on servers which are remote which are the control of users. It is a model which is
pay-per-use in that the structure benefactor by resources of service level
agreements(SLAs) which (Vaquero, Rodero-Merino, Caceres, & Lindner, 2009).
As cloud computing becomes prevalent, more and more sensitive information’s
are being centralized into the cloud. Such as emails, photo albums, personal
health records, financial transactions, tax documents and government documents
etc.

The detail that owners of data and cloud server are no extended in the similar
trusted domain may place the data at risk outsourced unencrypted. The cloud
server escape information of data to illegal enables or can be hacked. To deliver
privacy for data, data which is sensitive needs to be encrypted first of
outsourcing to the profitable public cloud (Kamara & Lauter, 2010). The
unimportant explanation of transferring all the information and decrypting in the
vicinity is clearly unreasonable, due to the gigantic amount of band width rate in
level of cloud scale systems.

Discovering preserving privacy and real search over encrypted date of cloud is
of supreme position seeing the possibly great amount of on claim users of data &
enormous quantity of outsourced document of data in the cloud, this problem is
predominantly challenging as it is tremendously difficult to meet also the
necessities of presentation, system usability and scalability encryption makes
operative utilization of data a very stimulating task given that there could be a
big quantity of outsourced information files. Also in the cloud computing owners
of data may portion their outsourced information with numerous users who
might want to only recover certain exact files of data. They are engrossed in
during a conference. One of the utmost prevalent imposts to do so is whole
keyword search method licenses users to intelligently recuperate files of interest.
Need for information retrieval is the most commonly occurring commission in
Need for information retrieval is the most commonly occurring commission in
cloud to the user to from server. Usually, cloud servers complete relevance result
ranking in instruction to make the exploration as earlier. Such ranked search
scheme allows users of data to find the most applicable info quickly, instead of
returning undistinguishable results. Ranked search can stylishly remove
unnecessary traffic for the network by distribution back only the greatest data for
relevance which is highly wanted in the “Pay-As-You-Use” paradigm for cloud.

For confidentiality shield, such process of ranking, the authors, must not escape
any keyword connected data. On the other side, to recuperate the result for
accuracy of search as the authors as to recover the searching for user experience,
it is also necessary for such ranking system to support multiple keywords search,
as single keyword search often yields far too coarse results. As a common
practice indicated by today’s the authors search engines (e.g., Google search),
data users may tend to provide a set of keywords instead of only one as the
indicator of their search interest to retrieve the most relevant data. And each
keyword in the search request can help narrow down the search result further.
“Coordinate matching” (Witten,Moffat & Bell,1999), i.e., as numerous matches
as conceivable, is a the authors-organized similarity amount among such
multikeyword meaning to improve the relevance of result, and has been
extensively used in the plaintext data retrieval (IR) community. Though, how to
put on it in the encrypted cloud data search scheme remains a very stimulating
assignment because of distinguishing privacy and security problems, counting
numerous strict supplies like the privacy of data, the privacy of index, privacy of
keyword, and many more.

Cloud computing is known as a substitute to old-style data technology and has

been progressively familiar as the greatest important rotating point in the
expansion of information technology due to its inherent sharing of resource and
maintenance of low characters. Cloud computing is also an internet based
computing model, where the information is shared, resources and software are
on condition that on other devices and computers on demand upon demand. This
allows the end user to admission the cloud computing possessions anytime any
platform such as a mobile computing platform, cell phone Clouds are big
puddles of effortlessly serviceable and available resources which are virtualized.
The information and the applications of software compulsory by the workforces
are not deposited on their own processers; instead they are deposited on remote
servers which will be under the control of other users. it is a pay-per- use model
in which the organization earner by resources of service level agreements
customized (SLAs) (Vaquero, Rodero-Merino, Caceres, & Lindner, 2009)
Security in Cloud

There are a portion of interests to tolerant Cloud Computing, there are also some
significant walls to receipt (Seung Hwan, Gelogo & Park, 2012). One of the
greatest major fences to acceptance is the security, surveyed by matters
concerning acquiescence, privacy and matters which are authorized. Since Cloud
Computing characterizes a comparatively new figuring model, there is an
enormous deal of ambiguity about how security at every level (network, host,
application, data levels, etc.) can be attained and in what way security of
application is stimulated to Cloud Computing. That indecision has dependably
controlled information managers to state that security is their number one
apprehension with Cloud Computing. Security anxieties recount to hazard areas
such as outside data storage, dependence on the internet which is public, absence
of control, integration and multitenancy with security is internal. Associated to
conservative technologies, cloud has many exact topographies, such as its
countless gauge and the detail that resources going to cloud providers are
completely disseminated, heterogeneous and totally virtualized. Conservative
security machineries such as Identity authentication, and authorization are no
longer enough for clouds in their current form. For of the cloud facilities replicas
working, the working replicas, and practices cast-off to allow services for cloud.
Cloud computing may present-day dissimilar dangers to suggestion than old-
style IT resolutions. Unfortunately, participating safety into these explanations is
frequently supposed as making them more inflexible.

Search in Encrypted Cloud Data

As Cloud Computing turn into extensive, more delicate data are being
transported into cloud, such as individual health records, emails, confidential
videos and images, data for business finance, documents for government, etc. As
per this i.e., storage their information into the cloud, the data owners can be
reassured from the problem of information storing space and preservation so to
like the on-demand high brilliance storage for data service (Reddy, 2013).
Though, the reality that information suppliers and cloud servers are not in the
alike reliable area may put the subcontracted information at danger. By way of
the cloud server can no extended be completely reliable in such an environment
for cloud since of a variety of reasons, they are: the cloud server may leakage
data to illegal things or it may be slashed. It tracks that delicate information
characteristically can be encrypted before outsourcing for data confidentiality
and fighting undesirable admissions. Though, encryption for data makes data
utilization effectiveness and efficiency a very challenging task given that there
could be a large amount of outsourced data files. Furthermore, in Cloud
Computing, data owners/provider may share their outsourced data with many
users. The individual users shall wish to only recover certain exact files of data
they are absorbed in through a given conference. One among the most
recognized customs is to specifically recover files complete keyword-based
search as a substitute of regaining all the files which are encrypted like before
which is totally unreasoning in cloud computing circumstances
(Khan,Wang,Kulsoom & Ullah, 2013) Like this keyword-based search technique
allows users to meaningfully regain files of awareness and has approximately
valuable in search of plaintext situations, such as Google search. Miserably,
encryption for information limits user’s capability to perform search for keyword
and subsequently makes the out-of-date plain text search methods not
appropriate for Cloud Computing.

Lately, the cloud computing pattern (Mell & T. Grance, 2011) is transforming
the establishments in method of effective their information mainly in the method
they accumulation, admittance and process information (Paillier,1999). As a
developing calculating pattern, for cloud computing, it interests many
establishments to correlated potential for cloud in relations of flexibility, cost-
efficiency and rid of managerial overhead. Cloud will originate valuable and
delicate data about the real data matters by detecting the variable data admission
designs even if is an information is encrypted (Capitani, Vimercati, Foresti, &
Samarati, 2012; Williams, Sion, & Carbunar, 2008). Most often, governments
characteristic their computational procedures in accrual to their data to the cloud.
The advantage of cloud is that the privacy and security matters in the cloud
which circumvents the productions to use those plunders. The information can
be encrypted earlier subcontracting to cloud when information is highly delicate.
When information is encrypted, regardless of the fundamental encryption
system, it is very interesting to accomplishment any information mining
responsibilities ever decrypting the information (Samanthula, Elmehdwi, &
Jiang, 2014).

Cloud computing is extended fantasized dream of computing as usefulness, from

cloud customs can at all store their information into cloud so to like on-demand
in height excellence claims and facilities from a public lake of configuring
computing possessions (Vaquero, Lodero-Merino, Caceres, & Lindner, 2009).
Its countless elasticity and financial investments are inspiring both persons and
initiatives to outsource their local composite information organization scheme
initiatives to outsource their local composite information organization scheme
into cloud. To shield information confidentiality and battle the authors come
y*admissions in the cloud and outside, delicate information, e.g., electronic mail,
private healthiness histories, snap scrapbooks, tax papers, economic
communications supposed to be encrypted by information proprietors before
subcontracting to the profitable public cloud (Kamara & Lauter, 2010) This
takes away the old-style information use service basis on plaintext keyword
searches. The unimportance of transferring all the information and decrypting
nearby is obviously unreasonable, due to the enormous quantity of bandwidth
price in cloud gage schemes. Furthermore, aside from removing the local storing
organization, storage information into serves for cloud no drive except they can
effortlessly have examined and used. Therefore, discovering preserving privacy
and real service for search done encrypted data for cloud is of supreme position.
Seeing the possibly big amount of on-demand information operators and
enormous quantity of subcontracted information forms in cloud, this problem is
predominantly stimulating as it is tremendously problematic to encounter also
the necessities of presentation, scheme scalability and usability. To encounter the
real information recovery essential, the big amount of forms request the server
for cloud to achieve consequence relevance ranking, in its place of recurring
undistinguishable consequences. Ranked search scheme allows information
operators to discovery the greatest applicable data rapidly, somewhat than
categorization finished every competition in the gratified group (Singhal, 2001).
Search based on rank can gracefully eradicate pointless traffic for network by
distribution posterior only the most applicable information, which is extremely
necessary in “pay-as-you use” cloud model. For protection, of privacy such
ranking process, though, should not escape any keyword associated data.
Respective, to recover the effect for accurateness of search to recover the
experience for user for going through, it is important for system for ranking to
provision many searches based on keywords, as sole search for keyword
frequently produces far too uneven consequences. As a mutual repetition
designated by todays search for the trains (e.g. Search for Google), data
administrators may grade to convey a standard of watchwords in its place of just
exceptional as the pointer of their enthusiasm of hunt to recoup the most relevant
data. What's more, exclusively watchword seek request is brilliant to help thin
discouraged the outcome for pursuit extra. “Arrange coordinating” (Witten,
Moffat, & Bell,1999), i.e., as various rivalries as likely, is an efficient correlation
sum among such multi-watchword intending to enhance the outcome pertinence,
and broadly utilized as a part of the plaintext data recovery (IR) people group.
Be that as it may, how to apply it in the scrambled cloud information seek
framework remains an extremely difficult assignment in view of characteristic
framework remains an extremely difficult assignment in view of characteristic
security and protection snags, including different strict prerequisites like the
information security, the list protection, the catchphrase security, and numerous
others. In the writing, searchable encryption (Song, Wagner, & Perrig, 2000;
Golle, Staddon, & Waters, 2004) is a useful strategy that regards scrambled
information as reports and permits a client to safely seek through a solitary
catchphrase and recover archives of intrigue. In any case, coordinate use of these
ways to deal with the protected extensive scale cloud information usage
framework would not be fundamentally reasonable, as they are created as crypto
primitives and can't suit such high administration level prerequisites like
framework ease of use, client look in understanding, and simple data revelation.
Some current plans have been proposed to bolster Boolean watchword look
(Golle, Staddon, 2004; Shen, Shi, & Waters, 2009) as an endeavor to enhance
the hunt adaptability, they are as yet not sufficient to give clients adequate
outcome positioning usefulness. Our initial work (Wang, Cao, Li, Ren, & Lou,
2010) has known about this issue, and given the authors for the safe positioned
seek over scrambled information issue yet just for inquiries comprising of a
solitary watchword. The most effective method to outline a proficient scrambled
information seek system that backings multi-catchphrase semantics without
security ruptures remains a testing open issue. In this chapter, surprisingly, the
authors characterize and take care of the issue of multi-watchword positioned
look over encoded cloud information (MRSE) while safeguarding strict
framework savvy security in the distributed computing worldview. Among
different multikey word semantics, the authors pick the proficient comparability
measure of “facilitate coordinating”, i.e., whatever number matches as would be
prudent, to catch the pertinence of information reports to the pursuit question.
The authors utilize “internal item comparability” i.e., the quantity of question
watchwords showing up in a record, to quantitatively assess such similitude
measure of that archive to the pursuit inquiry. Amid the file development, each
record is related with a double vector as a subindex where each piece speaks to
whether comparing catchphrase is contained in the report. The pursuit is likewise
depicted as a parallel vector where each piece implies whether comparing
watchword shows up in this hunt ask for, so the comparability could be precisely
measured by the inward result of the question vector with the information vector.
Be that as it may, straightforwardly outsourcing the information vector or the
question vector will damage the file protection or the pursuit security. To meet
the test of supporting such multi-catchphrase semantic without protection breaks,
the authors propose an essential thought for the MRSE utilizing secure internal
item calculation, which is adjusted from a safe k-closest neighbor (kNN)
procedure and afterward give two fundamentally enhanced MRSE plots in a he
procedure and afterward give two fundamentally enhanced MRSE plots in a he
authors ordered way to accomplish different stringent protection necessities in
two danger models with expanded assault abilities. Our commitments are
abridged as takes after:

1. For the first occasion when, the authors investigate the issue of multi
keyword positioned seek over encoded cloud information, and build up an
arrangement of strict protection necessities for such a safe cloud
information usage framework.
2. The authors propose two MRSE plans considering the similitude measure
of “organize coordinating” while at the same time meeting distinctive
protection prerequisites in two diverse danger models.
3. Thorough examination exploring protection and effectiveness assurances
of the proposed plans is given, and investigations on this present reality
dataset additionally demonstrate the proposed conspires undoubtedly
present low overhead on calculation and correspondence.

Distributed computing has been considered as another model of big business IT

foundation, which can compose enormous asset of registering, stockpiling and
applications, and the authors clients to appreciate omnipresent, helpful and on-
request arrange access to a mutual pool of configurable processing assets with
extraordinary effectiveness and negligible monetary overhead. Pulled in by these
engaging components, both people and undertakings are roused to outsource
their information to the cloud, rather than buying programming and equipment to
deal with the information themselves. Regardless of the different favourable
circumstances of cloud administrations, outsourcing touchy data, (for example,
messages, individual the authors being records, organization back information,
government reports, and so forth) to remote servers brings protection concerns.
The cloud specialist organizations (CSPs) that keep the information for clients
may get to clients' delicate data without approval. A general way to deal with
ensuring the information secrecy is to encode the information before
outsourcing. Notwithstanding, this will bring about an immense cost as far as
information ease of use. For instance, the current systems on watchword based
data recovery, which are generally utilized on the plaintext information, can't be
specifically connected on the scrambled information. Downloading every one of
the information from the cloud and unscramble locally is clearly unrealistic.
Keeping in mind the end goal to address the above issue, scientists have planned
some universally useful arrangements with completely homomorphic encryption
(Gentry, 2009) or unmindful RAMs (Goldreich & Ostrovsky,1996).
Nonetheless, these techniques are not common sense because of their high
computational overhead for both the cloud separate and client. More down to
computational overhead for both the cloud separate and client. More down to
earth extraordinary reason arrangements, for example, searchable encryption
(SE) plans have made commitments as far as effectiveness, usefulness and
security. Searchable encryption plans for the customer to store the encoded
information to the cloud and execute catchphrase seek over cipher text area. Up
until this point, plentiful works have been proposed under various risk models to
accomplish different hunt usefulness. For example, single catchphrase inquiry,
closeness look, multi-watchword Boolean pursuit, positioned seek, multi-
catchphrase positioned look, and so forth. Among them, multikey word
positioned look accomplishes increasingly consideration for its pragmatic
relevance. As of late, some dynamic plans have been proposed to bolster
embedding and erasing operations on record accumulation. These are critical
fills in as it is profoundly conceivable that the information proprietors need to
refresh their information on the cloud server. Be that as it may, few of the
dynamic plans bolster productive multikey word positioned look.

Distributed computing is one method for processing. Here the figuring assets are
shared by numerous clients. The advantages of cloud can be stretched out from
individual clients to associations. The information stockpiling in cloud is one
among them. The virtualization of equipment and programming assets in cloud
invalidates the money related venture for owning the information stockroom and
its upkeep. Many cloud stages like Google Drive, iCloud, SkyDrive, Amazon
S3, Dropbox and Microsoft Azure give stockpiling administrations.

Security and protection concerns have been the major challenges in distributed
computing. The equipment and programming security instruments like firewalls
and so forth have been utilized by cloud supplier. These arrangements are not
adequate to secure information in cloud from unapproved clients because of low
level of straightforwardness (Cloud Security Alliance, 2009). Since the cloud
client and the cloud supplier are in the diverse put stock in area, the outsourced
information might be presented to the vulnerabilities (Cloud Security Alliance,
2009; Ren, Wang, & Wang, 2012; Brinkman, 2007). In this manner, before
putting away the important information in cloud, the information should be
encoded (Kamara & Lauter, 2010). Information encryption guarantees the
information classification and trustworthiness. To save the information
protection the authors must outline a searchable calculation that takes a shot at
scrambled information (Wong,Cheung, Kao, & Mamoulis, 2009). Numerous
specialists have been adding to seeking on scrambled information. The hunt
systems might be single catchphrase look or multi watchword seek (C. Wang, N.
Cao, J. Li, K. Ren, & W. Lou, 2010). In gigantic database, the inquiry may bring
Cao, J. Li, K. Ren, & W. Lou, 2010). In gigantic database, the inquiry may bring
about many reports to be coordinated with watchwords. This causes trouble for a
cloud client to experience all archives and have generally applicable reports.
Look in view of positioning is another arrangement, wherein the reports are
positioned in view of their pertinence to the watchwords (Singhal, 2001).
Practical searchable encryption systems help the cloud clients particularly in
pay-as-you utilize show. The analysts consolidated the rank of archives with
numerous watchword pursuit to think of proficient financially suitable
searchable encryption strategies. In searchable encryption, related writing,
calculation time what's more, calculation overhead is the two most as often as
possible utilized parameters by the specialists in the space for dissecting the
execution of their plans. Calculation time (moreover called “running time”) is
the period required to play out a computational procedure for instance looking a
watchword, creating trapdoor and so forth. Calculation overhead is identified
with CPU usage regarding asset distribution measured in time. In this
examination work, the authors dissect the security issues in distributed storage
and propose the authors for the same. Our commitment can be compressed as
takes after:

1. Interestingly, the authors characterize the issue of secure positioned

watchword seek over scrambled cloud information, and give such as the
authors convention, which satisfies the protected positioned look usefulness
with no pertinence score data spillage against watchword protection.
2. Exhaustive security examination demonstrated that our deviated based
positioned searchable encryption plot utilizing CRSA and B-tree to be sure
appreciates “as-solid as possible”. security ensure contrasted with past
searchable symmetric encryption (SSE) plans.
3. Broad exploratory outcomes exhibit the adequacy and productivity of the
proposed arrangement.

Distributed computing has changed the way businesses approach IT, the authors
them to end up noticeably more nimble, present new plans of action, offer more
administrations, and trim down IT costs. Distributed computing innovations can
be executed in a wide assortment of designs, under different administration and
arrangement models, and can exist together with numerous advancements and
programming outline strategies. The distributed computing foundation keeps on
acknowledging unstable development. The authors, for security experts, the
cloud displays a tremendous quandary: How would you grasp the advantages of
the cloud while keeping up security controls over your associations' advantages?
It turns into an issue of adjust to decide if the expanded dangers are genuinely
It turns into an issue of adjust to decide if the expanded dangers are genuinely
justified regardless of the nimbleness and monetary advantages. Keeping up
control over the information is foremost to cloud achievement. 10 years prior,
undertaking information regularly lived in the association's physical framework,
all alone servers in the association's server farm, where one could isolate touchy
data in individual physical servers. Today, by virtualization and the cloud,
information might be under the association's intelligent control, yet physically
put away in foundation possessed and oversaw by an alternate element. This
move in charge is the main reason new methodologies and systems are required
to guarantee associations can keep up information security. At the point when an
outside party claims, controls, and oversees foundation and computational assets,
how might you be guaranteed that business or administrative information stays
private and secure, and that your association is shielded from harming
information breaks? This makes cloud information security fundamental.
Distributed computing the significance of Cloud Computing is expanding also, it
is accepting a developing thought in the logical also, mechanical groups. The
NIST (National Institute of Standards and Technology) proposed the
accompanying meaning of distributed computing:

Distributed computing is a demonstrate for the authors ring advantageous, on-

request arrange get to a common pool of configurable processing assets (e.g.,
systems, servers, stockpiling, applications, and administrations) that can be
quickly provisioned and discharged with negligible administration exertion or
specialist co-op communication. This cloud show advances accessibility (Mell &
Grance, 2012).

The cloud enhances joint effort, nimbleness, versatility, accessibility, capacity to

adjust to varieties as indicated by request, speed up advancement work, and
gives potential to cost diminishment through enhanced and effective processing.
Distributed computing consolidates various processing thoughts what's more,
advances, for example, Service Oriented Architecture (SOA), The authors 2.0,
virtualization and different advances with dependence on the Internet, supporting
regular business applications online through the authors programs to fulfil the
processing needs of clients, while their product and information are kept up on
the servers.

Cloud Delivery Models are:

• Private Cloud: Cloud framework is provisioned for use by a solitary

association that involves different inhabitants. Private mists might be
 worked on-or off-premises and are behind the organization firewall.
• Public Cloud: A cloud specialist co-op offers administrations to different
organizations, scholastic foundations, government offices, and different
associations with get to by means of the Internet.
• Hybrid Cloud: Hybrid mists join two cloud conveyance models that stay
novel as elements, yet they are bound together by innovation that the
authors information and application transportability. Cloud blasting is a
case of one way undertaking that utilize cross breed mists to adjust loads
amid pinnacle request periods.
• Community Cloud: Cloud foundation is provisioned for the selective
utilization of a group of client associations with shared figuring
prerequisites for example, security, strategy, and consistence.

The Service layers for these conveyance models are:

• Infrastructure as an Administration (IaaS): Cloud framework is the

accumulation of equipment and programming that the authors the
fundamental qualities of the cloud. IaaS permits clients to self-arrangement
these assets to run stages and applications.
• Platform as an Administration (PaaS): PaaS the author clients to adjust
legacy applications to a cloud situation or create cloud-mindful applications
utilizing programming dialects, administrations, libraries, and other
designer devices. Programming as an administration (SaaS) – Users can run
applications by means of numerous gadgets on cloud foundation.

SECURITY IN CLOUD
Although there is a considerable measure of advantages to receiving Distributed
computing, there are additionally some extensive hindrances to acknowledgment
(Seung Hwan, Gelogo & Park. 2012). A standout amongst the most significant
obstructions to selection is the security, trailed by issues in regards to
consistence, protection and approved matters. Since Cloud Processing speaks to
a moderately new registering model, there is an enormous arrangement of
vulnerability about how security by any stretch of the imagination levels
(arrange, have, application, information levels, and so forth.) can been
accomplished and how application security is moved to Cloud Figuring. That
vulnerability has reliably driven data administrators to express that security is
their number one worry with Cloud Computing. Security concerns identify with
hazard territories, for example, outside information stockpiling, reliance on
hazard territories, for example, outside information stockpiling, reliance on
general society the authors, absence of control, multitenancy what's more,
incorporation with inside security. Contrasted with traditional advances, cloud
has numerous elements, for example, its extraordinary scale and the reality that
assets having a place with cloud suppliers are totally disseminated,
heterogeneous and totally virtualized. Regular security components, for
example, character, validation, and approval are no sufficiently longer for mists
in their present shape. Considering the cloud benefit models utilized, the
operational models, and the philosophies used to the authors cloud
administrations, cloud registering may exhibit distinctive dangers to an
affiliation than customary IT arrangements. Unfortunately, incorporating
security into these arrangements is regularly seen as making them more
inflexible.

The distributed computing research group, especially the Cloud Security

Alliance, has perceived security issues in cloud. In its Top Threats to Cloud
Processing Report (Ver.1.0) (Top Threats to Cloud Computing Report
(Ver.1.0),2010), it recorded seven top dangers to distributed computing:

1. Mishandle and terrible utilization of distributed computing.

2. Uncertain application programming interfaces.
3. Pernicious insiders.
4. Shared innovation vulnerabilities.
5. Information misfortune or spillages.
6. Record, administration and movement capturing.
7. Obscure hazard profile.

ENCRYPTED CLOUD DATA

By doing as such i.e., putting away their information into the cloud, the
information proprietors/suppliers can be eased from the authors right of
information storage room and support to appreciate the on-request high
fabulousness information stockpiling administration (Reddy, 2013).
Notwithstanding, reality that information suppliers and cloud server are not in
the comparative trusted space may put the outsourced information at hazard, as
the cloud server may no longer be completely confided in such a cloud situation
considering several reasons, they are: the cloud server may spill data substance
to unapproved elements or it might be hacked. It takes after that delicate
information ordinarily ought to be scrambled preceding outsourcing for
information ordinarily ought to be scrambled preceding outsourcing for
information security and battling undesirable gets to. Information encryption
makes information usage adequacy and proficiency an extremely difficult
assignment given that there could be a lot of outsourced information records.
Moreover, in Cloud Computing, information proprietors/supplier may impart
their outsourced information to a substantial number of clients. The individual
clients may be earning to just recover certain exact information records they are
occupied with all through a given session. A standout amongst the most
acknowledged routes is to specifically recover documents through catchphrase
based hunt as an option of recovering all the scrambled documents back which is
totally outlandish in distributed computing situations (Khan & Wang, 2010).

BACKGROUND
Firstly, they use the “Dormant Semantic Analysis” to uncover relationship
amongst terms and reports. The inert semantic examination exploits certain
higher-arrange structure in the relationship of terms with archives (“semantic
structure”) and receives a diminished measurement vector space to speak to
words and reports. In this manner, the relationship be the authors terms is
naturally caught. Furthermore, their plan utilizes secure “k-closest neighbour (k-
NN)” to accomplish secure hunt usefulness. The proposed plan could return the
correct coordinating records, as the authors as the documents including the terms
inactive semantically related to the question watchword. At long last, the
exploratory result exhibits that their strategy is superior to the first MRSE
conspire (Song & Wagner, 2000). Their procedures have various pivotal focal
points. They are provably secure: they give provable mystery to encryption, as in
the untrusted server can't learn anything about the plain content when just given
the cipher text; they give inquiry disengagement to quests, implying that the
untrusted server can't learn much else about the plaintext than the query output;
they give controlled seeking, so that the untrusted server can't hunt down a
subjective word without the client's approval; they likewise bolster concealed
inquiries, so that the client may approach the untrusted server to hunt down a
mystery word without uncovering the word to the serve (Sun, Wang, Cao, Li,
Lou, Hou, & Li, 2013). They propose a tree-based list structure and different
versatile techniques for multidimensional (MD) calculation so that the handy
hunt proficiency is greatly improved than that of direct inquiry. To further
improve the inquiry protection, they propose two secure file plans to meet the
stringent protection prerequisites under solid danger models, i.e., known cipher
text display and known foundation demonstrate. What's more, they devise a plan
upon the proposed file tree structure to the authors genuineness check over the
returned indexed lists. At long last, the authors show the viability and
productivity of the proposed plots through broad trial assessment. (Yu, Lu, Zhu,
Xue, & Li, 2013) Distributed computing has developing as a promising example
for information outsourcing and astounding information administrations.
Worries of touchy data on cloud conceivably causes security issues. Information
encryption ensures information security to some degree, yet at the cost of traded
off proficiency. Searchable symmetric encryption (SSE) permits recovery of
encoded information over cloud. In this chapter, the concentration is on tending
to information security issues utilizing SSE. Interestingly, the authors define the
security issue from the part of comparability significance and plan vigor. They
watch that server-side positioning considering request saving encryption (OPE)
spills information security. To take out the spillage, the authors propose a two-
round searchable encryption (TRSE) conspire that backings beat k multikey
word recovery. (Wong, D, Cheung, Kao, & Mamoulis, 2009) In This chapter
they talk about the general issue of secure calculation on an encoded database
and propose a SCONEDB (Secure Computation ON an Encrypted Database)
show, which catches the execution and security necessities. As a contextual
analysis, the authors concentrate on the issue of k-closest neighbour (kNN)
calculation on an encoded database. The authors build up another lopsided
scalar-item protecting encryption (ASPE) that jelly a unique kind of scalar item.
They utilize APSE to develop two secure plans that bolster kNN calculation on
scrambled information; each of these plans is appeared to oppose down to earth
assaults of an alternate foundation learning level, at an alternate overhead cost.
Broad execution studies are done to assess the overhead and the proficiency of
the plans (Zhang & Zhang, 2011). Since Boneh et al. proposed the thought and
development of Public Key Encryption with Keyword Search (PEKS) conspire,
numerous updates and expansions have been given. Conjunctive watchword
inquiry is one of these expansions. A large portion of these built plans cannot
understand conjunctive with subset catchphrases look work. Subset watchwords
look implies that the beneficiary could inquiry the subset catchphrases of all the
catchphrases implanted in the cipher text. The authors ponder the issue of
conjunctive with subset watchwords seek work, talk about the disadvantages
about the existed plans, and after that give out a more effective development of
Public Key Encryption with Conjunctive-Subset Keywords Search (PECSK)
conspire. A correlation with different plans about effectiveness will be
displayed. They additionally list the security prerequisites of their plan, then give
out the security investigation (Song, Wagner, & Perrig, 2000).
MAIN FOCUS OF THE AUTHORS

Issues, Controversies, Problems

Data Owner

The information proprietor subcontracts her information to the cloud for suitable
and consistent data admission to the equivalent search operators. To defend the
information confidentiality, the information proprietor encrypts the unique
information over symmetric encryption. To recover the exploration
effectiveness, the data owner makes approximately keywords for each
subcontracted document. The equivalent index is then formed giving to the
keywords and a secret key. Afterward, the data owner directs the encrypted
brochures and the equivalent directories to the cloud, and directs the symmetric
key and secret key to exploration workers.

Cloud Server

The cloud server is an in-the authors entity which supplies the encrypted
brochures and equivalent indexes that are established from the information
proprietor, and delivers information admittance and exploration facilities to
exploration users. When an exploration operator directs a keyword access to the
cloud server, it resolves re-emergence a group of equivalent brochures founded
on firm processes.

Search User

A search operator enquiries the subcontracted brochures from the cloud server
with subsequent three steps. First, the exploration operator accepts together the
symmetric key and the secret key from the information proprietor. Second, as
per the exploration keywords, the exploration operator usages the secret key to
make hatch and directs it to cloud server. Last, he obtains the corresponding text
collection from the cloud server and is decrypting them with the symmetric key.

RSA Algorithm

RSA is the process shaped by the modern computers to decrypt and encrypt
messages. It is an asymmetric cryptographic technique. Asymmetric resources
messages. It is an asymmetric cryptographic technique. Asymmetric resources
that there are two different keys. This is public key cryptography, since one of
them can be prearranged to everyone. The other key must be private. In
conclusion, the factors of an integer are rigid (the factoring problem). A worker
of RSA kinds and then issues the produce of two big prime figures, with a
supplementary rate, as public key of theirs. The factors which are prime remain
secret. Anybody can use the public key to encrypt a communication, but with
presently published approaches, if the public key is large, only somebody with
information of the prime factors can practicably decode the communication. It is
used by modern computers to encrypt and decrypt communications. It is an
asymmetric cryptographic procedure. Asymmetric means that there are two
different keys. This is public key cryptography, since one of them can be given
to all. The other key must be kept private.

Hierarchical Clustering

It produces hierarchy of clusters.

Two approaches:

1. Agglomerative:

Being a bottom-up approach. Individually opinion starts in its individual cluster.

And couple of clusters combine as one changes up the hierarchy.

2. Divisive:

It is top-down approach. Explanations starts in one cluster, and splits are

achieved recursively as one transfers up the hierarchy.

SOLUTIONS AND RECOMMENDATIONS

Symmetric Key Algorithm

1. Symmetric encryption practices the similar key to mutually decrypt and

encrypt
2. DES is used as symmetric key algorithm
3. They are fast and their complexity is low so they can be easily
implemented
4. Secret key should be configured in all hosts
 4. Secret key should be configured in all hosts

Asymmetric Key Algorithm

1. Asymmetric key uses one key to decrypt and one key to encrypt
2. RSA is one of the asymmetric key algorithm
3. It is slow then symmetric key algorithm
4. No need of configuring secret keys in all hosts

Knn Algorithm

1. Partitional clustering
2. Partitions independent of each other
3. Sensitive to cluster center initialization
4. Poor convergence speed and bad overall clustering can happen due to
poor initialization
5. Works only for around shapes
6. Doesn’t work well for non-convex shapes

Hierarchical Clustering Algorithm

1. Hierarchical clustering
2. Visualize using a tree structure
3. Can give different partitioning
4. Doesn’t need specification of number of clusters
5. It can be slow
6. Two types agglomerative and divisive

FUTURE RESEARCH DIRECTIONS

The authors propose to additionally spread the application to deliberate the
extensibility of the set of file and the multiuser environments of cloud. This way,
some initial consequences on the extensibility and the multiuser cloud
environments. Another stimulating theme is to progress the highly ascendable
searchable encryption to allow the authors organized exploration on large
practical databases.
CONCLUSION
Fine-grained multikey word search (FMS) method was studied over encrypted
data from cloud and delivered two FMS schemes. The FMS I constitute both the
preference factors and relevance scores of keywords to improve more accurate
search and improved users’ knowledge, correspondingly. The FMS II attains
safe and the authors organized search with functionality, i.e., “AND”, “OR” and
“NO” operations of keywords. Additionally, the authors have projected the
improved methods supporting classified sub-dictionaries (FMSCS) to advance
efficiency. The authors have used hierarchical clustering computation for the
same.

REFERENCES
Abdalla, M., Bellare, M., Catalano, D., Kiltz, E., Kohno, T., Lange,
T., & Shi, H. (2008). Searchable encryption revisited: Consistency
properties, relation to anonymous ibe, and extensions . Journal of
Cryptology , 21(3), 350–391. doi:10.1007/s00145-007-9006-6

Alliance, C. S. (2009). Security Guidance for Critical Areas of Focus

in Cloud Computing. Retrieved from
http://www.cloudsecurityalliance.org

BallardL.KamaraS.MonroseF. (2005). Achieving efficient conjunctive

keyword searches over encrypted data. Proc. of ICICS.
10.1007/11602897_35

BellareM.BoldyrevaA.NeillA. O. (2007). Deterministic and efficiently

searchable encryption. Proc. of CRYPTO.

Boldyreva, A., Chenette, N., Lee, Y., & Oneill, A. (2009). Order-
preserving symmetric encryption . In Advances in Cryptology-
EUROCRYPT (pp. 224–241). Springer.
Boneh, Kushilevitz, Ostrovsky, & W. E. S. III. (2007). Public key
encryption that allows pir queries. Proc. of CRYPTO.

BonehD.CrescenzoG. D.OstrovskyR.PersianoG. (2004). Public key

encryption with keyword search. Proc. of EUROCRYPT.

BonehD.WatersB. (2007). Conjunctive, subset, and range queries on

encrypted data. Proc. of TCC, 535–554.

Brinkman, R. (2007). Searching in encrypted data. PhD thesis.

Brinkman. (2007). Searching in encrypted data. PhD thesis.

Cao, N., Wang, C., Li, M., Ren, K., & Lou, W. (2014). Privacy-
preserving multikeyword ranked search over encrypted cloud data .
IEEE Transactions on Parallel and Distributed Systems , 25(1), 222–
233. doi:10.1109/TPDS.2013.45

Cao, N., Wang, C., Li, M., Ren, K., & Lou, W. (2014). Privacy-
preserving multikeyword ranked search over encrypted cloud data .
IEEE Transactions on Parallel and Distributed Systems , 25(1), 222–
233. doi:10.1109/TPDS.2013.45

Cao, Wang, & Li, Ren, & Lou. (2014). Privacy-preserving

multikeyword ranked search over encrypted cloud data. IEEE
Transactions on Parallel and Distributed Systems , 25(1), 222–233.

Cash, J., & Jutla, K., Ro3u, & Steiner. (2013). Highly-scalable
searchable symmetric encryption with support for Boolean queries.
Proc. CRYPTO, 353-373.

ChangY.-C.MitzenmacherM. (2005). Privacy preserving keyword

searches on remote encrypted data. Proc. of ACNS.
10.1007/11496137_30
CurtmolaR.GarayJ. A.KamaraS.OstrovskyR. (2006). Searchable
symmetric encryption: improved definitions and efficient
constructions. Proc. of ACM CCS. 10.1145/1180405.1180417

De Capitani di Vimercati, S., Foresti, S., & Samarati, P. (2012).

Managing and accessing data in the cloud: Privacy risks and
approaches. CRiSIS, 1 –9.

Dinh, Lee, Niyato, & Wang. (2013). A survey of mobile cloud

computing: Architecture, applications, and approaches. Wireless
Commun. Mobile Comput., 13(18).

Gentry, C. (2009). A fully homomorphic encryption scheme (Ph.D.

dissertation). Stanford University.

Goh, E.-J. (2003). Secure indexes. Retrieved from

http://eprint.iacr.org/2003/216

Goldreich, O., & Ostrovsky, R. (1996). Software protection and

simulation on oblivious rams . Journal of the Association for
Computing Machinery , 43(3), 431–473. doi:10.1145/233551.233553

GolleP.StaddonJ.WatersB. (2004). Secure conjunctive keyword

search over encrypted data. Proc. of ACNS, 31–45. 10.1007/978-3-
540-24852-1_3

Hwan, Gelogo, & Park. (2012). Next Generation Cloud Computing

Issues and Solutions. International Journal of Control and
Automation, 5.

Hwang, Y., & Lee, P. (2007). Public key encryption with conjunctive
keyword search and its extension to a multiuser system . Pairing.
doi:10.1007/978-3-540-73489-5_2

JungT.MaoX.LiX.TangS.-J.GongW.ZhangL. (2013). Privacy

JungT.MaoX.LiX.TangS.-J.GongW.ZhangL. (2013). Privacy
preserving data aggregation without secure channel: multivariate
polynomial evaluation. Proceedings of INFOCOM, 2634–2642.
10.1109/INFCOM.2013.6567071

Kamara & Lauter. (2010). Cryptographic cloud storage. In RLCPS.

Springer.

Kamara, S., & Lauter, K. (2010). Cryptographic cloud storage. In

RLCPS. Springer. doi:10.1007/978-3-642-14992-4_13

Kamara, S., & Lauter, K. (2010). Cryptographic cloud storage. In

RLCPS. Springer. doi:10.1007/978-3-642-14992-4_13

Katz, J., Sahai, A., & Waters, B. (2008). Predicate encryption

supporting disjunctions, polynomial equations, and inner products .
Proc. of EUROCRYPT. doi:10.1007/978-3-540-78967-3_9

Khan, Wang, Kulsoom, & Ullah. (2013). Searching Encrypted Data

on Cloud. International Journal of Computer Science Issues, 10(6).

Khan, Wang, Kulsoom, & Ullah. (2013). Searching Encrypted Data

on Cloud. International Journal of Computer Science Issues, 10(6).

LewkoA.OkamotoT.SahaiA.TakashimaK.WatersB. (2010). Fully

secure functional encryption: Attribute-based encryption and
(hierarchical) inner product encryption. Proc. of EUROCRYPT.
10.1007/978-3-642-13190-5_4

Li, H., Dai, Y., Tian, L., & Yang, H. (2009). Identity-based
authentication for cloud computing. In Cloud Computing. Berlin,
Germany: Springer-Verlag. doi:10.1007/978-3-642-10665-1_14

Li, H., Liu, D., Dai, Y., Luan, T. H., & Shen, X. (2014). Enabling
efficient multikeyword ranked search over encrypted cloud data
through blind storage . IEEE Transactions on Emerging Topics in
Computing . doi:doi:10.1109/TETC.2014.2371239

Li, J., Wang, Q., Wang, C., Cao, N., Ren, K., & Lou, W. (2010).
Fuzzy keyword search over encrypted data in cloud computing. Proc.
of IEEE INFOCOM’10 Mini-Conference.
doi:10.1109/INFCOM.2010.5462196

Li, R., Xu, Z., Kang, W., Yow, K. C., & Xu, C.-Z. (2014). Efficient
multikeyword ranked query over encrypted data in cloud computing .
Future Generation Computer Systems , 30, 179–190.
doi:10.1016/j.future.2013.06.029

Liang, Cai, Huang, Shen, & Peng. (2012). An SMDPbased service

model for interdomain resource allocation in mobile cloud networks.
IEEE Trans. Veh. Technol., 61(5).

Liang, H., Cai, L. X., Huang, D., Shen, X., & Peng, D. (2012). An
smdpbased service model for interdomain resource allocation in
mobile cloud networks . IEEE Transactions on Vehicular Technology
, 61(5), 2222–2232. doi:10.1109/TVT.2012.2194748

Mahmoud & Shen. (2012). A cloud-based scheme for protecting

source-location privacy against hotspot-locating attack in wireless
sensor networks. IEEE Trans. Parallel Distrib. Syst., 23(10).

Mahmoud, M. M., & Shen, X. (2012). A cloud-based scheme for

protecting source-location privacy against hotspot-locating attack in
wireless sensor networks . IEEE Transactions on Parallel and
Distributed Systems , 23(10), 1805–1818.
doi:10.1109/TPDS.2011.302

Mell & Grance. (2011). The nist definition of cloud computing (draft).
NIST Special Publication, 800, 145.
Naveed, Prabhakaran, & Gunter. (2014). Dynamic searchable
encryption via blind storage. Proceedings - IEEE Symposium on
Security and Privacy , 639–654.

Paillier, P. (1999). Public key cryptosystems based on composite

degree residuosity classes. Eurocrypt, 223–238. doi:10.1007/3-540-
48910-X_16

Pang, Shen, & Krishnan. (n.d.). Privacy-preserving similarity-based

text retrieval. ACM Transactions on Internet Technology, 10(1), 4.

Reddy. (2013). Techniques for Efficient Keyword Search in Cloud

Computing. International Journal of Computer Science and
Information Technologies, 4(1).

Ren, K., Wang, C., & Wang, Q. (2012). Security Challenges for the
Public Cloud . IEEE Internet Computing , 16(1), 69–73.
doi:10.1109/MIC.2012.14

Samanthula, B. K., Elmehdwi, Y., & Jiang, W. (2014). k-nearest

neighbor classification over semantically secure encrypted relational
data. eprint arXiv:1403.5001

ShenE.ShiE.WatersB. (2009). Predicate privacy in encryption

systems. Proc. of TCC.

Shen, Q., Liang, X., Shen, X., Lin, X., & Luo, H. (2014). Exploiting
geodistributed clouds for e-health monitoring system with minimum
service delay and privacy preservation . IEEE Journal of Biomedical
and Health Informatics , 18(2), 430–439.
doi:10.1109/JBHI.2013.2292829

Shen, Liang, Shen, Lin, & Luo. (2014). Exploiting geodistributed

clouds for a e-health monitoring system with minimum service delay
and privacy preservation. IEEE J. Biomed. Health Inform., 18(2).

Singhal, A. (2001). Modern information retrieval: A brief overview .

A Quarterly Bulletin of the Computer Society of the IEEE Technical
Committee on Data Engineering , 24(4), 35–43.

Singhal, A. (2001). Modern information retrieval: A brief overview .

A Quarterly Bulletin of the Computer Society of the IEEE Technical
Committee on Data Engineering , 24(4), 35–43.

Singhal, A. (2001). Modern information retrieval: A brief overview .

A Quarterly Bulletin of the Computer Society of the IEEE Technical
Committee on Data Engineering , 24(4), 35–43.

SongD.WagnerD.PerrigA. (2000). Practical techniques for searches

on encrypted data. Proc. of S&P.

Song, D. X., Wagner, D., & Perrig, A. (2000). Practical techniques for
searches on encrypted data. In Proceedings of S&P. IEEE.

Song, D. X., Wagner, D., & Perrig, A. (2000). Practical techniques for
searches on encrypted data. Proceedings of S&P, 44–55.

Stefanov, Papamanthou, & Shi. (2014). Practical dynamic searchable

encryption with small leakage. Proc. NDSS.
doi:10.1109/TPDS.2013.282

Sun, Wang, Cao, Li, Lou, Hou, & Li. (2013). Verifiable privacy-
preserving multikeyword text search in the cloud supporting
similarity-based ranking. IEEE Transactions on Parallel and
Distributed Systems. DOI: 10.1109/TPDS.2013.282

Sun, W. (2013). Privacy-preserving multikeyword text search in the

cloud supporting similarity-based ranking. Proc. 8th ACM SIGSAC
Symp.Inf., Comput. Commun. Secur., 71-82.
Symp.Inf., Comput. Commun. Secur., 71-82.

Sun, Wang, Cao, Li, Lou, Hou, & Li. (2013). Verifiable privacy-
preserving multikeyword text search in the cloud supporting
similarity-based ranking. IEEE Transactions on Parallel and
Distributed Systems. DOI: 10.1109/TPDS.2013.282

Vaquero, L. M., Rodero-Merino, L., Caceres, J., & Lindner, M.

(2009). A break in the clouds: Towards a cloud definition . ACM
SIGCOMM Comput. Commun. Rev. , 39(1), 50–55.
doi:10.1145/1496091.1496100

Vaquero, L. M., Rodero-Merino, L., Caceres, J., & Lindner, M.

(2009). A break in the clouds: Towards a cloud definition . ACM
SIGCOMM Comput. Commun. Rev. , 39(1), 50–55.
doi:10.1145/1496091.1496100

Vaquero, L. M., Rodero-Merino, L., Caceres, J., & Lindner, M.

(2009). A break in the clouds: Towards a cloud definition . ACM
SIGCOMM Comput.Commun. Rev. , 39(1), 50–55.
doi:10.1145/1496091.1496100

WangC. (2010). Secure Ranked Keyword Search Over Encrypted

Cloud Data. Proc. ICDCS ’10. 10.1109/ICDCS.2010.34

WangC.CaoN.LiJ.RenK.LouW. (2010). Secure ranked keyword

search over encrypted cloud data. In Proceedings of ICDCS. IEEE.
10.1109/ICDCS.2010.34

WangC.CaoN.LiJ.RenK.LouW. (2010). Secure ranked keyword

search over encrypted cloud data. Proc. of ICDCS’10.
10.1109/ICDCS.2010.34

Wang, Yu, Lou, & Hou. (2014). Privacy-preserving multikeyword

fuzzy search over encrypted data in the cloud. Proceedings - IEEE
INFOCOM .

Williams, P., Sion, R., & Carbunar, B. (2008). Building castles out of
mud: practical access pattern privacy and correctness on untrusted
storage. ACM CCS, 139–148. doi:10.1145/1455770.1455790

Witten, Moffat, & Bell. (1999). Managing Gigabytes: Compressing

and indexing documents and images. Morgan Kaufmann Publishing.

Witten, Moffat, & Bell. (1999). Managing gigabytes: Compressing

and indexing documents and images. Morgan Kaufmann Publishing.

Wong, W. K., Cheung, D. W., Kao, B., & Mamoulis, N. (2010).

Secure kNN computation on encrypted databases. Proc. ACM
SIGMOD Int. Conf. Manage. Data, 139-152.

Wong, W. K., Cheung, D. W., Kao, B., & Mamoulis, N. (2009).

Secure knn computation on encrypted databases. Proc. of SIGMOD.
10.1145/1559845.1559862

WongW. K.CheungD. W.-l.KaoB.MamoulisN. (2009). Secure knn

computation on encrypted databases. Proceedings of SIGMOD
International Conference on Management of Data, 139–152.
10.1145/1559845.1559862

Yang, L., Liu, & Yang. (2014). Secure dynamic searchable symmetric
encryption with constant document update cost. Proc.GLOBECOM.

Yang, Li, Liu, Yang, & M. (2014). Secure dynamic searchable

symmetric encryption with constant document update cost. In
Proceedings of GLOBCOM. IEEE.

Yu, J., Lu, P., Zhu, Y., Xue, G., & Li, M. (2013). Towards secure
multikeyword top-k retrieval over encrypted cloud data . IEEE
Transactions on Dependable and Secure Computing , 10(4), 239–250.
doi:10.1109/TDSC.2013.9

Yu, J., Lu, P., Zhu, Y., Xue, G., & Li, M. (2013). Towards secure
multikeyword top-k retrieval over encrypted cloud data . IEEE
Transactions on Dependable and Secure Computing , 10(4), 239–250.
doi:10.1109/TDSC.2013.9

Zhang, B., & Zhang, F. (2011). An efficient public key encryption

with conjunctive-subset keywords search . Journal of Network and
Computer Applications , 34(1), 262–267.
doi:10.1016/j.jnca.2010.07.007

KEY TERMS AND DEFINITIONS

Big Data: Large data sets that are analyzed computationally.

Cipher-Text: Encrypted form of text.

Cloud Computing: Provides shared computing resources.

Clusters: Unit of allocable hard disk space.

Encryption: Process of converting information into code.

Hierarchical Clustering: Which builds a hierarchy of clusters.

Minimum Hash Subtree: Root is having the minimum value in the subtree.

Symmetric Scheme: Algorithm used for cryptography.

 CHAPTER 6
Research Analysis of Development Pipelines in Augmented and
Virtual Reality Technologies

Pronay Peddiraju
VIT University, India

P. Swarnalatha
VIT University, India

ABSTRACT
The purpose of this chapter is to observe the 3D asset development and product
development process for creating real-world solutions using augmented and
virtual reality technologies. To do this, the authors create simulative software
solutions that can be used in assisting corporations with training activities. The
method involves using augmented reality (AR) and virtual reality (VR) training
tools to cut costs. By applying AR and VR technologies for training purposes, a
cost reduction can be observed. The application of AR and VR technologies can
help in using smartphones, high performance computers, head mounted displays
(HMDs), and other such technologies to provide solutions via simulative
environments. By implementing a good UX (user experience), the solutions can
be seen to cause improvements in training, reduce on-site training risks and cut
costs rapidly. By creating 3D simulations driven by engine mechanics, the
applications for AR and VR technologies are vast ranging from purely computer
science oriented applications such as data and process simulations to mechanical
equipment and environmental simulations. This can help users further
familiarize with potential scenarios.

INTRODUCTION

Background of Study

Augmented and Virtual Reality allows a user to experience computer generated

Augmented and Virtual Reality allows a user to experience computer generated
environments and simulations in an immersive perspective view using head
mounted devices (HMDs). This project aims at shedding light at some of the
development pipelines involved in creating Augmented and Virtual Reality
products. By leveraging software aimed at creating these simulative
environments involving 3D assets and computer-generated levels, these
Augmented Reality (AR) and Virtual Reality (VR) products can be used to
provide solutions to numerous real world problems. These solutions range from
environment simulators to training software that can be used by corporates and
individuals to reduce costs and improve efficiency. The various pipelines used in
creating these products include 3D asset development, Level Design using a
Game Engine, implementation of required audio and acoustics, creating the
required deliverable for the target platform and the final implementation in the
operating environment.

The scope for AR and VR technologies can be seen in numerous fields ranging
from game development to creating training simulations that can be used in the
industry to cut operating costs as well as visualize final products in real time.
Industries that can benefit from these technologies however will see the
requirement of a heavy initial investment depending on the complexity and the
demand of the product created. Typically AR applications are less resource
hungry and can be used on smartphones and have a lower initial investment
while most VR applications are bulky, resource hungry and need powerful
computing environments paired with precise HMDs hence making it a more
expensive yet smoother experience.

This thesis aims at understanding the development pipelines involved in making

AR and VR solutions and providing a development model catered to the same.
From a software engineering perspective, there are numerous models that are
applicable for use in the AR and VR domain but to obtain the best possible
result, it is important to have a model that is specific for these types of solutions.
By conducting a thorough study on the pipelines involves, we can derive some
conclusions that can help us understand the requirement of resources for each
and also identify the workflow in the process. The workflow can then be further
simplified by implementing a parallel model while developing the various
components involved in the process.

PROBLEM STATEMENT
Augmented and Virtual Reality have great scope when applied onto mobile
Augmented and Virtual Reality have great scope when applied onto mobile
device hardware as it can reach out to a larger user base (Jason Jong Kyu Park,
2013). But due to the complex nature of the development process and the
technology itself, it can be difficult to identify a solid model that can be used to
develop, maintain and re-iterate if required on the software. Engine
optimizations continue to be made to allow more precise colours, details,
tracking and operation (Jacob B. Madsen, 2014). This allows the development
process to continue to get simpler with respect to using a game engine but there
is no method that correlates the work flow between each of the pipelines. The
evolution of VR and AR is extremely fast paced and is witnessing the extinction
of numerous software add-ons (Dey, 2016). There is a requirement to be able to
modify and re-iterate on existing products in order to ensure the scalability and
longevity of the said product. There are numerous restrictions on the polygon
count of the 3D models that can be successfully rendered in a VR environment
due to restrictions on the hardware capabilities of hand held devices (Ahmad
Hoirul Basori, 2015). This is to be identified properly and accounted for before
the start of the development process to ensure that no issues with respect to
support over the platform is encountered. It is important to create a smart UI
(User Interaction) that is both aware of position and orientation of the user in
order to provide a better user experience within the application (Hsin-Kai Wu,
2103). A common problem faced in the making of VR and AR applications is
providing a good UI and UX (User Experience).

By conducting a review of the problems identified above, we can account for

each and propose a model that can identify with the VR and AR development
pipeline making sure that each of the pipelines involved can be handled in a way
to ensure minimization of the issues listed in the above paragraph.

MOTIVATION
The motivation behind this study is to understand the development process
revolving around VR and AR technologies and leveraging the tools in an
efficient manner. By creating VR and AR based simulations, we are able to
create a more realistic 3D perspective that the users can find a lot more
comfortable to orient themselves to. By the virtue of this study, we will try to
compare and contrast various methodologies for the application development
process implementing VR and AR and eventually try to conceptualize a model
for the same.
From a software engineering point of view, it is difficult to categorise VR and
AR applications and implement a specific model for development due to its
complex and flexible nature. To be able to visualize the development process
and come to conclusions on the time line and nature of the project, it is important
that research on the same is performed to implement a more efficient
methodology or follow a model designed specifically for these applications.

By further understanding the working principles of VR and AR, we can derive a

specific model for the development of the same. The model will account for all
the development processes involved and hence provide a holistic view of the
application development.

DEVELOPMENT PIPELINES
Although there is no specific rules in developing for VR and AR applications,
the most common approach is to make use of a Game Engine paired with some
3D modelling software. The common approach revolves around the following
processes:

• Product Conceptualization: The process of understanding requirements

of the product. Analysis on its operating environment. Planning the required
resources that will be involved. Implementing the required SDK for the
target platform. Understanding the scope of the project and its future
support.
• Asset Development (2D or 3D): The process of developing the required
assets that will be used in the application. This involves creating 2D
images, 3D object and even the generated animations, textures and other
such elements that go into the game engine.
• Audio Engineering: The process where the required audio is created with
respect to its use case and application. The audio is designed and exported
in a format that can be implemented on the game engine.
• Level Design: The process of designing the 3D level that the user will be
a part of. This involves placement of assets in the vicinity of the user and
implementing the required assets at the required places within the
application in a 3D space.
• Implementing Required SDKs: Linking the required SDK within the
project to allow references to be made within the application. The SDK will
provide features specific to the technology or platform that will be in use.
 • Programming Application Logic: The process of linking all the assets
and the level with logic that governs their working in the application.
Usually this process goes hand-in-hand with the level design process. The
application logic will define the working principles of every element in the
application and can range from anywhere between 1 simple script to several
scripts that are applied on a variety of scene elements.
• Analysis of Operating Environment: The analysis of the operating
environment to account for specific requirements that may arise for the
implementation of the product in the said environment. In this phase of the
development, the required changes will be made to the application in order
to allow support for the target environment.
• Testing on Target Platform: This is a vital process in the development
cycle. Upon development of the software, the testing phase identifies any
issues with the application. Testing on the target platform will allow the
developers to interact with application and identify and changes that are to
be made in order to have the smooth functioning of the product before
deployment.
• Deployment: The phase that involves shipping of the product to the target
user base or client. The deployment phase requires the packaging of the
project and ensuring an easy installation onto the target platform. It should
account for any cases that may not allow a smooth installation and
implement changes accordingly to make the process complete successfully.
• Support: The post deployment phase that ensure that any bugs or issues
encountered can be dealt with by taking the required steps and making the
necessary changes to the product. It allows the developers to roll out
updates and fixes that maintains the smooth working of the product by
increasing both the efficiency and longevity.

Each of these development processes will be discussed more elaborately in the

coming sections of this document. There may be additional steps involved in
development of VR and AR applications besides those listed above. For the
purpose of our research, we will only be accounting for the most commonly used
approach which will implement the above steps in most cases.

In this document, we will observe the steps involved in each process, the study
of what tasks may be performed parallel to others so as to improve efficiency,
methods that can optimize the mentioned pipeline and devise a model that can be
implemented for the same. Since all the integration and use of assets is
performed in the Game Engine that is implemented for development, the focus
will primarily be on how to improve the process with the Game Engine in
will primarily be on how to improve the process with the Game Engine in
perspective.

NEED FOR DEVELOPMENT MODEL

By taking apart each of the pipelines for development, we can get a solid
understanding of the steps involved in making assets and integrating them with
the product. The study will apply this concept in understanding which of the
pipelines can be worked upon simultaneously and which of the pipelines need to
be given a greater priority.

The model shall provide an insight to the application flow and how to create a
progress methodology that reduces risk, cost and other potential elements that
can cause drawbacks in the development process. At the same time, the model
will also provide a framework that can be implemented to improve efficiency
and performance over the target platform by performing the required software
planning.

To come to conclusion about the above mentioned process and derive a model
that can be used, we will be implementing a VR application to understand the
same. The proposed system shall be applying the use of VR on a smartphone and
implement the use of 3D objects as well as audio within the application.

Introduction to Related Concepts

The proposed system is applying the use of a simple hand-held smartphone

combined with a low cost software solution that may be implemented by using
the said smartphone for AR and VR applications. By leveraging the Unity3D
Game Engine features and combine those with those provided by the GoogleVR
SDK for unity, we are able to create a variety of 3D simulative spaces that can
be both interactive and immersive. The control that user has over the application
is restricted to a single button interaction and the user point-of-view based visual
selection.

Figure 1 is a screenshot providing a more concrete view of the GoogleVR

Reticle Pointer being used in the scene to create point of view based selections.

Figure 1. VR scene view

VR Camera Setup

In order to create a realistic 3D perspective view, we require the use of 2

cameras in the designed scene. As can be seen in Figure 1, the VR view provides
a render using both the left eye and the right eye camera which is later merged
into 1 single perspective view that provides information on both lighting and
depth. This allows the user to experience a very realistic environment.

VR Reticle Pointer

The VR Reticle pointer is used to provide the user a point on the centre of the
screen so that the user can identify where his point-of-view is and what objects
within his view are intractable. As can be observed from Figure 1, the user
reticle (white circle on the centre of the screen) will scale up and form a circle
when there is an intractable object on the scene before the user. In this case we
see the teleportation cylinder being highlighted to green when the user looks
towards it.

VR View Port

The view port for VR implements the use of the 2 cameras as discussed. The 2
cameras provide 2 views on the view port that will create a single image when
used in a VR headset (Due to the human perspective). The view port does not
used in a VR headset (Due to the human perspective). The view port does not
only account for the generated views but also provides depth, contrast and
provides realistic imagery based on scene lighting.

Lighting

Lighting is among the most important elements within the level designed. If the
lighting within the scene is not done properly, it could affect the look of the level
and hence affect the way the objects on the scene are perceived by the user. With
low quality lighting, the objects on the scene will not provide the same levels of
realism which in turn creates a poor user experience within the application.

User Interface

The user interface in case of VR is slightly different compared to the regular UI

implemented while developing games using a game engine. The UI for VR or
AR applications must rely on world space and not create an overlay on the
screen which is not ideal for all scenarios. The UI must also be placed in a way
that it allows the user to have a clear view of it within the game scene. A good
practice is to ensure that the UI is always facing the user irrespective of his
orientation. This will allow the user to view the UI with greater clarity in world
space.

LITERATURE REVIEW
From the mentioned reference documents, we are able to obtain an evolutionary
view on the VR and AR development processes over the years. Some of the gaps
identified from these documents are listed as follows

1. Interaction with VR and AR technologies is still restricted in most cost

effective implementation cases. Due to the complexity of the product and
the development process, its implementation is usually restricted if the
initial investment is low. This can be countered with better planning and
use of an optimized model that can allow cost cutting measure allowing for
investment in additional technologies that can help improve the user
interaction.
2. The use of VR and AR is still not common as the initial investment for
the use of these technologies is relatively high. Again an issue that is
evident due to the complexity of the development process and cost of
 evident due to the complexity of the development process and cost of
technology required to implement the product. This cannot necessarily be
countered by implementation of a development model but the model may
create scope for implementing a solution that can reduce the investment
making it a more cost friendly implementation.
3. There are many limitations in terms of hardware that restrict the ability of
developers to create VR and AR based solutions. This requires the need for
optimization and efficiency. The optimizations that are to be made should
account for the operating environment and by proposing a valid model, the
planning and development can be done by making note of the same.
4. VR may not always seem to be a practical solution as level complexity
on the engine increases. This makes it harder for a developer to identify and
modify the scenes and levels created within the game engine. By using a
model that implements a proper planning phase and implementing a
reference document alongside the development process, the above gap can
be accounted for.
5. There are limitations with use of physics and applying forces on objects
due to the orientation and conditions created by using the headset. Users
encounter psychological stresses and may even experience nausea as the
VR developer has the power to simulate a world before the viewer. The
addition of forces in the scene or manipulating the camera in a way that
affects the user’s position should hence be avoided. These conditions can be
implemented by compiling a valid development methodology.
6. There is still no proper solution for the problem of nausea and some other
psychological effects observed by AR and VR users after using the
application over long durations. Although steps can be taken to avoid this
problem by implementing a valid development methodology, due to the
subjective nature of this problem, sufficient data must be obtained before
development of the product. This will allow the developers to account for
the various results obtained on how and why these experiences are
observed. Hence making a more reliable and usable software.

ARCHITECTURE IMPLEMENTED
Figure 2 is the block diagram describing the architecture of the working model
used to perform analysis on various aspects of the pipeline.

Figure 2. Proposed system architecture.

Figure 2 describes the architecture of the Augmented Reality platform that will
be used to analyse the steps and processes of all the pipelines involved in its
development.

Given below is a brief description of the functions of each of the subsystems

present in the proposed system architecture diagram:

• Application Subsystem: The system that is the chunk of code that defines
the application and its working methodologies combined with a user
interface for operation.
• Context Subsystem: The subsystem that collects context data from
various subsystems. Examples include preferences and progress data.
• World Model Subsystem: The subsystem that controls the use of real
world elements in the 3D world space within the application.
Implementation of this subsystem is the major difference between
Augmented Reality and Virtual Reality based solutions.
• Tracking Subsystem: This subsystem used the sensors present on the
device to track the user’s position with respect to the 3D space. Tracking is
the module that provides smooth video feed and movement flow in the
application during run time.
 • Interaction Subsystem: The subsystem that controls the user input and
output. All forms of interaction between the user and the system is
controlled by this subsystem.
• Presentation Subsystem: This subsystem is responsible for how the
application is generated and presented to the user. It includes modules such
as the render engine and texture repositories.

IN-DEPTH ANALYSIS
To understand the 6 subsystems in the proposed model we shall go through their
respective functional diagrams. The functioning and interaction between each
subsystem is integral in ensuring the smooth flow of the application and to
ensure that nothing within the application is asynchronous at any given point.

Application Subsystem

To understand the application subsystem, we will refer to Figure 3 in order to

observe the links and flow within the system. In this subsystem the code
interacts with the various libraries and referenced game objects to provide
meaning to the implementations. There may be numerous co-routines, singletons
and other such entities within the application’s source code for controlling a
chunk of the application. An example of a singleton would be the use of a game
manager code script that can only exist as a single instance within the scene
designed.

The code will interact with both the Engine editor and Engine headers in order to
provide functionality within the scene. Additionally the code will also work
alongside the used SDKs and Plugins such as iTween (a commonly used tool for
providing game object based transitions and animations). The references to the
SDKs provide the required SDK features within the application such as the use
of the VR Reticle and the dual Scene Camera setup.

Context Subsystem

The Context Subsystem communicates with all the subsystems to put together
context settings and details for the application including the preferences,
progress data and the in app setting that have been setup by the user. This
subsystem performs a rough system analysis to understand the working
principles and regulates it to a recommended setting to ensure the smooth flow
principles and regulates it to a recommended setting to ensure the smooth flow
of the application.

The context files interact with internal system details and read pre-sets that exist
within the hardware to be able to optimize working of the application for the
specific working environment at hand.

World Model Subsystem

The world model subsystem is primarily used as a manager for the level
designed in the application. Ensuring the working of the level correctly and
maintaining the proper perspective vision within the application is handled by
the World Model Subsystem. Within this subsystem, we account for real time
motion tracking accounting for the world model view with respect to the
rotation, orientation and movement of the user by using the smartphone platform
on a head mounted device (HMD).

By implementing the link to the Tracker module in the device, the World Model
Subsystem can be able to pick up reading from the tracker to provide an accurate
update to the world view and ensure a stable performance in real time as any
movement performed in the real world will be immediately reflected in the
virtual world.

Tracking Subsystem

The Tracking subsystem is responsible for the tracking of the user by making
use of the available sensors on the HMD or device used with the HMD. For
example, in the case of the VR headsets such as Oculus Rift or HTC Vive, the
tracking will be performed using some specialized sensors that are part of the
HMD to ensure smoothness and maintaining a real time 110 degrees field of
view. While in the case of the smartphone (the case in our application for the
experiment), we observe that the smartphone sensors such as the gyroscopic
sensor and in some cases the availability of special hardware such as Tango by
Google can be implemented for spatial tracking within the application.

This subsystem mainly interacts with sensors and ensure that the feedback
received from them is accurately accounted for by the remaining subsystems.

Interaction Subsystem
This subsystem uses the available input devices to provide interaction between
user and application. In the case of VR and AR the interactive elements in the
scene may be in great numbers but to allow user to interact with the application
we have limited hardware capabilities. This is why the interaction subsystem is
used to control the variety of permutations that are performed to maximize the
utility of available inputs.

This subsystem can only interact with a limited set of inputs but these inputs are
manipulated by conversion to individual events. The event system hence
generated is used to provide the required functionalities and interaction for the
user.

Presentation Subsystem

This subsystem is complex in nature due to the tasks it is responsible for

handling. It provides the application presentation for the user by loading up the
required textures and by making use of a rendering engine. This requires the
most computational power as it generates the level that the user interacts with.
Figure 3 is a sample image of a level that is generated within the Game Engine.

Figure 3. Presentation Subsystem

Figure 3 shows the render of a scene within the engine taken from the
perspectives of both clients in the scene. The clients require a generated view of
perspectives of both clients in the scene. The clients require a generated view of
their own and both are required to be generated real time. The characters on the
scene and the weapons on the scene are rendered by this subsystem and provides
a real-time output which in this case is synchronized via a network.

PROPOSED DESIGN

Overview

To further understand the development cycle from a software development

perspective, the proposed model is to explore the following in depth to come to a
conclusive analysis of the process:

The methodology implemented to create AR or VR applications involves:

1. Using the Unity Game engine to create the environment (either 3D or

2D).
2. Using a 3D asset development software such as Autodesk Maya for
creating 3D models and animations that can be incorporated in the
environment designed on Unity 3D.
3. Performing the required SDK integration. For the case of this project, it
was to link the Unity application created with the android SDK to be able to
run on any ARM based Android device (not limited to ARM).
4. Setting up the run time environment on the desired device. In the case of
this project that involved getting android setup on a smartphone and making
the run environment support the generated Unity game in terms of texture
resolution support and memory management.
5. Finally it was integrating the application on other android devices to
experience the AR environment and interact with the physical world and
the AR level generated through the application.

For the purpose of this project, the methodology will incorporate the use of a
level created using the Unity 3D engine. By using 3D models, audio and other
such assets the scene will be created to provide a realistic user experience. To
incorporate the integration of various smartphone features to improve the
experience, the project will be exported to support the format of the device
(Windows OS for PC, Mac OS for Apple platforms, Android for Android based
smartphones, iOS for iPhones, etc). The final process will be creating a build
that will run as an application on the target platform and provide an interface to
share data between the operating system and the VR platform in real time to
share data between the operating system and the VR platform in real time to
provide a rich user experience.

EXPECTED RESULT
To be able to create a Virtual Reality application that runs on an android device
and provides interaction with the real world and the generated level to the user.
The application will highlight the pipelines for developing Virtual Reality
applications and provide a rich and immersive user experience to solve a real-
world problem.

The application will use a series of technologies to provide a solution to the

immersive interaction experience. These technologies include spatial tracking,
GPS co-ordinate based VR tracking, and real-world perspective using the
generated VR level. The result of developing this application should hence
provide further clarity on how integration of VR applications with the regular
operating system can be performed to provide a clear and immersive user
experience. This project’s scope reaches out to provide a rich user experience
while using software solutions by making them user friendly, interactive,
simplistic and be cheaper alternatives to existing platforms.

IMPLEMENTATION

Application Design

Mentioned below is the implementation that is being developed as part of this

project. To understand the working principles of the 3D development pipeline,
the below mentioned work includes details from the asset development phase to
the level design and application programming phase.

VR Level Design

Figure 4. Use of a 3D asset on a VR capable level.

In Figure 4, we can see the use of a 3D asset on the scene the view of which in
game can be found in the Game tab of the unity 3D software. The scene is a
simple construction of a base plane and 6 assets that are used as teleportation
points in the game. The teleportation points work as intractable objects that will
make the user teleport to the selected location. The 3D asset was made using
Autodesk 3DS Max software and shall be made intractable by implementing
animations that will be trigger via the user’s VR reticle. This implementation
displays the level design involved to some extent in a VR application.

Game View VR

Figure 5. Game View and reticle pointer for the generated VR level
The game view (Figure 5) shows the use of the reticle and interactive elements
that have their light turn green on hover while they remain red when they are not
interacted with by the reticle. This is an example of programmed logic that is
being performed on each of the interactive element that defines its working
principles within the application. The level designed applies materials on the
objects depending on the interaction from the user. In case of no selection
performed, the object is assigned a material that provides the red glow. Once the
object has been hovered over, it applies a material that provides the green glow
while the reticle is still above the intractable object. After the user has hovered
over the intractable element and the reticle in no longer on the object, a blue
material is assigned providing a blue glow that signals that the user has seen the
object and the reticle has hovered over it at some point in the application.

OPERATING ENVIRONMENT ANALYSIS

Below mentioned are all the hardware and software components used for
developing the mentioned application:

Hardware

• Android Device: Any android device that adheres to the minimum

specification requirements in order to run applications developed on the
 Unity 3D engine.
• Augmented Reality Headset:The headset will carry your smartphone
with a pair of lenses that provide an AR experience. Examples of such
headsets include GearVR, VR Box and Google Cardboard.
• Raspberry Pi (To provide a controller for the user):If the application
requires some form of user interaction, a raspberry pi can be used to
provide the interaction with user by implementing IR sensors.

Software

• Unity 3D Software: Unity 3D is a Game Engine that can be used to

develop applications that incorporate either 2D or 3D world space. This
engine will be used for the purpose of this project to develop an android
based scene that can provide an AR experience while using it in Google
Cardboard.
• Autodesk Maya: Autodesk Maya is a 3D modelling software provided by
Autodesk. It will be used in this project to provide all the 3D assets and
animations that will be incorporated in the projects.
• Android Studio: Used to provide the integration of the C# based
application code with the java based android platform by making use of the
intents provided by android.

ACKNOWLEDGMENT
The author would like to thank the School of Computer Science and
Engineering, VIT University and special thanks to Dean SCOPE, for his kind
guidance and support along with our guide Dr. Prabu S and Dr. Swarnalatha P,
without whom it would not have been possible to complete this mammoth task.
Also, I would like to thank the anonymous reviewers and the editor-in-chief of
the IGI global publishers for their valuable guidance which has improved the
quality and presentation of the chapter.

REFERENCES
Basori, Afif, Almazyad, Abujabal, Rehman, & Alkawaz. (2015). Fast
Markerless Tracking for Augmented Reality in Planar Environment.
3D Research, 6(4), 1-11.

DeyA. (2016). A Systematic Review of Usability Studies in

Augmented Reality between 2005 and 2014. Mixed and Augmented
Reality (ISMAR-Adjunct), 2016 IEEE International Symposium on.

Jason, J. K. P., Park, Y., & Mahlke, S. (2013). Efficient execution of

augmented reality applications on mobile programmable accelerators.
Field-Programmable Technology (FPT) 2013 International
Conference on, 176-183.

Kristensen, B. B., May, D., & Nowack, P. (2011). Collaboration and

Modeling in Ambient Systems: Vision Concepts and Experiments.
System Sciences (HICSS) 2011 44th Hawaii International Conference
on, 1-6.

MacWilliams, Reicher, Klinker, & Bruegge. (2014). Design Patterns

for Augmented Reality Systems. CEUR Workshop Proceedings , 19.

Madsen, J. B., & Stenholt, R. (2014). How wrong can you be:
Perception of static orientation errors in mixed reality. 3D User
Interfaces (3DUI) 2014 IEEE Symposium on, 83-90.
10.1109/3DUI.2014.6798847

Teather & Stuerzlinger. (2016). SIVARG: Spatial Interaction in

Virtual/Augmented Reality and Games. Proceedings of the 2016.
doi:10.1109/ISMAR-Adjunct.2016.0036

Wu, H.-K., Lee, S. W.-Y., Chang, H.-Y., & Liang, J.-C. (2013,
March). Current status, opportunities and challenges of augmented
reality in education . Computers & Education , 62(C), 41–49.
doi:10.1016/j.compedu.2012.10.024
 CHAPTER 7

Iterative MapReduce:
i-MapReduce on Medical Dataset Using Hadoop
Utkarsh Srivastava
VIT University, India

Ramanathan L.
VIT University, India

ABSTRACT
Diabetes Mellitus has turned into a noteworthy general wellbeing issue in India.
Most recent measurements on diabetes uncover that 63 million individuals in
India are experiencing diabetes, and this figure is probably going to go up to 80
million by 2025. Given the rise of big data as a socio-technical phenomenon,
there are various complications in analyzing big data and its related data
handling issues. This chapter examines Hadoop, an open source structure that
permits the disseminated handling for huge datasets on group of PCs and thus
finally produces better results with the deployment of Iterative MapReduce. The
goal of this chapter is to dissect and extricate the enhanced performance of data
analysis in distributed environment. Iterative MapReduce (i-MapReduce) plays a
major role in optimizing the analytics performance. Implementation is done on
Cloudera Hadoop introduced on top of Hortonworks Data Platform (HDP)
Sandbox.

INTRODUCTION
We live in the age of data. It’s not easy to measure the total volume of data. IDC
estimate puts the size of the “digital universe” at 4.4 zettabytes in 2013 and is
forecasting a tenfold growth by 2020 to 44 zettabytes. Such enormous volumes
forecasting a tenfold growth by 2020 to 44 zettabytes. Such enormous volumes
of data suffer from various issues like storage capabilities and synchronization
problems since they are stored at different places depending upon the vicinity to
data servers. Given the rise of Big Data as a socio-technical phenomenon there
are various complications in analyzing Bigdata and its related data handling
issues. In such a case Iterative MapReduce comes in really handy. The term
'Enormous Data' alludes to the monstrous volumes of both organized and
unstructured information which can’t be directly utilized with conventional
database administration frameworks.

With the quick increment in the diabetic patients in India and number of
determinants for the diabetes, the information will become tremendous and turns
out to be Big Data which couldn't be handled by traditional DBMS. Here we
discuss about Hadoop, an open source structure that permits the disseminated
handling for huge datasets on group of PCs and thus finally produces better
results with the deployment of Iterative MapReduce. The main goal is to dissect
and extricate the enhanced performance of Data Analysis in distributed
environment. Iterative MapReduce (i-MapReduce) plays a major role in
optimizing the analytics performance. Implementation is done on Cloudera
Hadoop introduced on top of Hortonworks Data Platform (HDP) Sandbox.
Hortonworks Hadoop is used for the extraction of useful data patterns in light of
the inquiry identified with different determinants of diabetes dataset obtained
from IBM quest. Iterative MapReduce algorithm for sequential pattern mining
utilizes a distributed computing environment. It consists of two processes a
mapping process and reducing process which is further utilized in two separate
phases namely Scanning phase and Mining phase. During the scanning phase
high performance is gained by distributing the task of finding elements over
different mapper tasks which can be run parallel on multiple machines with a
distributed database or file system. In the mining phase the mapper task creates a
lexical sequence tree for finding patterns and to improve efficiency of limited
depth dfs which is run on the tree and the reducer task finds the support value for
the patterns and thus in turn finds the useful patterns. So, to avoid the problem of
the serialized processing, we opt for parallel processing in Big Data
environment. This parallel processing not only reduces the computation time but
also optimizes the resource utilization.

This is a brief idea about the chapter and we will see its implementation and
impacts in further sections.
BACKGROUND
Big Data is also like normal data but with an enormous size. This is a term is
generally used to describe a collection of data that is very huge in size and still
growing exponentially with time. In short, such a large collection of data which
is difficult to handle via traditional databases and other management tools is
called ‘BigData’. Generic features of BigData are:

• Volume
• Velocity
• Variability
• Veracity

The main objective of this analysis is to find interesting patterns on the basis of
conditional dependence on given attributes of a dataset. With such an increased
rate of data generation it becomes very difficult to analyze the patterns in the
dataset. Also the situation becomes more critical when we have sequential
patterns in the dataset i.e. the order of dependency matters. Such behavior of
data is very usual in day to day actions such as customer shopping behavior,
medical symptoms leading to a future patient disease, financial stock market data
predictions etc. Pattern mining of BigData using Hadoop faces a lot of issues in
terms of data storage, data shuffling, data scanning, data processing units etc.

Figure 1. Implications of BigData

Sequential pattern mining is one of the most important data mining technique
used in various application technologies in modern world. Some examples are
Gene Analysis, Intrusion detection of System attack and Customer Behaviour
Prediction. The centralized logic behind sequential pattern mining is to find
frequent sequences within a transactional or operational database. The formal
definition can be detailed as follows. Definition 1: Let D be a sequence database,
and I = {y1, · · ·, ym} be a set of m different items. S = {d1, · · ·, di} is a
sequence consisting of an ordered list of itemsets. An itemset di is a subset of
items ⊆ I. A sequence Dr = {r1, · · ·, rn} is a subsequence of sequence Dt = {t1, ·
· ·, tm} where 1 ≤ i1 < · · · < in ≤ m such that r1 ⊆ ti1, r2 ⊆ ti2, · · ·, rn ⊆ tin.
Sequential pattern mining is generally used to find all the sets of sequential
patterns whose occurrence frequencies ≥ minimum support ∗ |D|. Here
minimum support is the support threshold value. Some of the patterns which will
be of great help in this regard are Apriori-based, Pattern growth-based and
Projection-based algorithms.

Pattern mining algorithms are mainly of two types:

• Join based algorithm

• Tree based algorithm

Join based algorithm includes basic Apriori algorithm which is executed in

recursive level-wise manner. It includes three major steps that are repeatedly
executed again and again for different values of k (the length of the pattern).
First step involves the generation of candidate patterns followed by the pruning
of patterns and finally the validation of patterns against the given dataset.
Apriori algorithm finds the enum tree by making use of bfs (breadth first search).
It generates the candidate key using join-based approach. Observing the
execution of Apriori, we can see that enum tree is not explicitly used but the tree
after scanning phase is constructed on the basis of dataset prefixes. Other major
algorithms in this field are TreeProjection and FP-growths. Both of these
algorithms make use of hierarchical relations between the given dataset and the
useful patterns between them. It reduces the number of data scans and thus
having lesser resource requirements.

Tree based algorithm includes DHP (Direct Hashing and Pruning) which was
proposed as an improvisation for Apriori algorithm. It is built on traditional logic
only but with two major improvisations. The first improvisation is to prune the
itemsets in each iteration, and the second improvisation is to trim the
transactions so that support-counting becomes more efficient. After first scan of
the dataset it maintains a DHT (Distributed Hash Table) and uses it for all
the dataset it maintains a DHT (Distributed Hash Table) and uses it for all
further calculations. It is this DHT which is used for making a BitMap
representation table for determining the presence or absence of an item set in the
dataset. For presence of an item in the dataset its normalized frequency count is
taken into consideration with all conditional attributes.

IBM’s Watson Analytics, introduced initially for the finance industry, attempts
to provide a simple analysis system based won IBM’s sophisticated Cogno’s
processing capabilities. It uses NLP to give predictive decisions based on the
input data. It was initiated with an objective to help business processes to access
data remotely for coming up with business related decisions. All types of
businesses generate data these days. It is done by means of websites, social
media, shopping patterns of customers, user experience, etc. This reflects the
need for companies to strategize on various aspects of storage and mining of
useful information for the available data. This is considerably more challenging
than just locating, identifying, understanding, and citing data. In order to take
decisions based on the available data, the systems require being fully fledged
automated which can store data and schedule tasks at regular time intervals to
phase out results from time-time.

MAIN FOCUS OF THE CHAPTER

The main agenda of this chapter is to understand about traditional pattern mining
algorithms and inculcating improvisations in the same for BigData processing in
a distributed environment. We will see how data is stored in master and slave
nodes and how they can be synchronized to operate interchangeably for
improved results. This methodology will not only reduce the computation time
but will also improve resource utilization. The best solution to this situation is
using parallel computing along with MapReduce algorithms in a distributed
environment. It involves a lot of parameters such as support count, confidence,
discrete attributes, continuous attributes, leveling tags etc. On the basis of all
these attributes, patterns can be classified as:

• Negative Frequent Patterns

One of the major challenges in pattern mining in the segregation of useful

patterns from the others. It is this reason because of which different attributes are
taken into consideration to come up with proper segregation. Some of them are
support, confidence, normalization factor etc. Suppose an item has some
frequency of occurrence in a dataset but this does not gives us a real picture of
the item. There are many conditional parameters which determines the
occurrence of the item. So we can say that absolute frequency of an item has no
meaning in pattern mining. It needs to be conditional considering all the biases
and interrelated factors. Various pattern mining approaches focus on frequency
normalization and they get the correct picture of patterns in a dataset. In case of
negative association, the general observation is that the occurrence of one item
with greater support means the absence of other item in the dataset.

• Constrained Frequent Patterns

Most of the times we need to filter the patterns according to user requirements
and make an appropriate choice of the distinguishing parameters. For example a
client may want the presence of one specific itemset in the mining dataset but
defines some differentiating factors which determine the conditional rule base
for that particular mining pattern. We can in fact directly push all the constraints
in the mining process. It has several advantages as mining can be performed at
much lower support levels as compared to other methods. Thus we can say that
constrained pattern mining involves deployment of conditional patterns with
normalized frequency values.

• Compressed Frequent Patterns

One of major issues in pattern mining is the volume and redundancy of patterns
in a given dataset. All traditional algorithms need to make multiple data scans
for finding interesting patterns. This increases the operational and system
requirements. Most of the times a subset of a frequent pattern set is also
frequent. Also to make the process easy, most algorithms make use of bitmap
representations of the useful patterns and simultaneously pruning the remaining
patterns. Therefore it solves the problem of multiple data scans and also reduces
the computational requirements.

The main objective of the proposed system is to make this process hassle free for
the users i.e. they will not have to maintain clusters and spend huge amounts for
the hardware stuff to mine information as well as schedule tasks responsible for
Big Data analysis. This system is built upon the Hadoop ecosystem. It will be a
mix of standalone and hosted service for the clients and will greatly reduce their
workload. It aims at finding all interesting patterns in the dataset based on the
user input and thus helps in categorizing the useful patterns from the others. It
aims at reducing the hardware requirements for the computation and produce
resource optimized results. It will greatly boost the development of any product
or services in the market industry with reference to:

• Investors
• Business Analyst
• E-commerce companies
• Advertising agencies
• Wholesaler

If the transactional database has huge amounts of frequent patterns then it

becomes difficult to load complete data in the main memory. So to overcome
this problem various parallel mining techniques can be used which parallely
mines the frequent patterns and gives the desired output. Now to solve this
problem we can make use of Sequential Pattern Mining on the Cloud which
makes use of cloud framework. It deploys bitmap representation to find the
mining patterns. Traditional approaches operates in centralized fashion which
has restricted computation power and efficiency. Sequential Pattern Mining on
the Cloud provides a good solution for this problem. It makes use of recursive
MapReduce techniques with two major phases. Firstly it scans the dataset and
then mines the dataset. During scanning phase it scans the data and makes a
distributed hash table out of it. Now in the mining phase it makes use of iterative
MapReduce processes which forms simultaneous Mappers and executes the
process concurrently. It deploys a distributed version of sequential pattern
mining which utilizes the divide and conquer strategy. Our algorithm utilizes
MapReduce computing algorithm for sequential pattern mining in a distributed
environment. It consists of two processes a mapping process and reducing
process which is further utilized in two separate phases namely Scanning phase
and Mining phase.

In the scanning phase high performance is gained by distributing the task of

finding elements over different mapper tasks which can be run parallel on
multiple machines with a distributed database or file system. Reducer tasks are
then run to find the frequency of different patterns and infrequent patterns we
subsequently removed from the records. The result can be stored to be passed on
to the mining phase.

In the mining phase the main aim is to find all possible patterns in the dataset
and to prune away non-sequential and non-useful patterns to filter out the
important information from the dataset. We try to achieve this using a lexical
important information from the dataset. We try to achieve this using a lexical
sequence tree which can generate all possible patterns existing in the dataset and
then this tree can be pruned away to find out useful or sequential patterns.

So now comes a very obvious question, how is iterative MapReduce better than
traditional algorithms in distributed environment. Here are some of the
comparing parameters:

• Memory Scalability: This method greatly reduces the memory

requirements because of Master-Slave architecture. Here the dynamic
nature of data is also taken into consideration which helps is parallel
processing. The required memory is also reduced in a way that all the
unnecessary patterns are continuously being pruned off.
• Work Partitioning: Most of the data processing jobs are divided in two
phases namely the Scanning phase and the Mining phase. Both these phases
have a Mapper task and a Reducer task. Now the execution of mapper
phase of the master will not hamper the execution of mapper phase of the
slave nodes and vice-versa. This will eliminate the mutual dependency of
various MapReduce tasks.
• Scaling: Setting up Hadoop is a one-time investment and after that it can
be easily scaled up with new cluster nodes. It can be deployed to
economical commodity servers, permitting the groups of nodes to develop
as required.
• Data Restrictions: Hadoop can handle unstructured data without the need
to process it before use. There is no need to massage the data before it can
be used. So, MapReduce is preferred slightly more here.
• Language: The language behind MapReduce's control component is
basically Java. It also makes use of internal Pig and Python which will
make the process of management and deployment easier.

ISSUES, CONTROVERSIES, AND PROBLEM

With such an increasing rate of Data generation it is becoming very difficult to
store all the data on a single device and compute it locally. This is the reason that
Distributed computing comes into picture. The major issue with traditional
pattern mining algorithms is that it needs to scan the same dataset again and
again thus increasing the computational requirements. Also there is no discrete
method of segregating the useful patterns from the uninteresting patterns. Most
of these algorithms make use of local computation which has the constraint of
storing data locally and having repeated data scans. Another application lies in
the field of genome sequencing. Genome sequencing decodes the order of DNA
nucleotides, or bases, in a genome—the order of A’s, C’s, G’s, and T’s that
make up an organism's DNA. The human genome is constituted of over 3 billion
of these genetic letters. Samples are collected from the patients suffering from
genetic disorders, genome sequencing is carried out to determine the appropriate
genetic abnormality in patients. These genome sequencing samples generate
terabytes of data which need to analyze in real time to provide accurate results to
decipher and verify with existing sequences. There are many more places of
applications of data analysis in the domain of Big Data. There are many such
scenarios where we require instant results from large datasets but due to various
constraints we are not able to get results on time. These may be due lack of
processing power, lack of skillset etc. Thus Iterative MapReduce(i-MapReduce)
comes to our rescue and reduces the dependancy on Hardware modules and
clusters.

So to overcome all these issues we can make use of distributed version of

sequential pattern mining which utilizes the divide and conquer strategy. The
main focus of the algorithm is to utilize MapReduce computing algorithm for
sequential pattern mining in a distributed environment. It consists of two
processes a mapping process and reducing process which is further utilized in
two separate phases namely Scanning phase and Mining phase.

SOLUTIONS AND RECOMMENDATIONS

The architecture comprises of the following elements Data Storage, Data parser,
Data sampler, Data analysis and finally data visualization layer.

The sampling layer is responsible for collecting random pieces of data in order
to perform a few initial preprocessing tasks. It performs the task of selecting a
subset from the original set of all measurements. Since unstructured data is fed
into this layer so the task of structuring it is performed by the sampling layer.
This is done by means of finding the delimiter of the input data and finding
useful summarizations of the sampled data to get an overview of the data to run
optimized algorithms on the same. The types of sampling ranges from simple
random sampling which in which any particular data has equal probability of
being sampled to custom sampling where the user provided how much
percentage of data is to be sampled from the input data.
The data parsing layer supports extraction of various attributes of any type of
data. This layer supports in various formats like CSV,XML and JSON since
most of the data meant for analysis is available in any one of these formats. This
layer also supports extraction of specific attributes of the data. This is again
based on custom queries as provided by the user for any of the three input types.

The core analysis layer is subdivided into various sublayers. These sublayers are
named aggeration, filtering, classification and data organisation layers
respectively. The data enters into this layer after being sampled and parsed and
hence is ready for the type of analysis which has been specified by the user. The
objective of the filtering layer is to extract certain chunks of data on the basis of
some applied function or randomly extracting chunks which vary in size and
dimension. Major functionality in this layer is finding all distinct values present
in a string of parameters so that all repetitions are filtered out thus reducing
redundancy of patterns.

The objective of the visualization layer is that it helps in getting a better

understanding of the data. This layer makes the picture clearer to a general user.
After running a specified algorithm on the input data set, an output file will be
generated which will be fed into the software named Tableau along with the
cluster details and port number. This will generate a graphical representation of
our output data for improved understanding.

It has two main phases:

1. Scanning Phase

During the scanning phase high performance is gained by distributing the task of
finding elements over different mapper tasks which can be run parallel on
multiple machines with a distributed database or file system. Reducer tasks are
then run to find the frequency of different patterns and infrequent patterns we
subsequently removed from the records. The result can be stored to be passed on
to the mining phase. The major task in this phase is to create the lexical sequence
tree. Each common node or its associated subtree represents a sequence in the
data set. We gain scalability in this phase using by splitting the data set into parts
and running a mapper on each independently to offer better execution time on
distributed systems. The mapper runs a depth search on the tree mine the
patterns but the search is a depth limited search to reduce the load on the mapper
program. This removes a bottleneck in the LST construction and mining and
gives more balance to each mapper phase. The mapper phase creates an
gives more balance to each mapper phase. The mapper phase creates an
intermediate output which can be processed by the reducer processed to prune
away non-sequential patterns. The pattern is stored in the format<pattern, bit-
AND -result> for the reducer side to process. The mapper phase outputs the
patterns along with the node depth and a threshold value is assumed for the node
depth any node which has support count lower than the threshold is not
processed and its extensions are also not processed.

Pseudo Code for Scanning Phase (Mapper)

Client Input: p (partitioner)

Step 1: Variable p = to take input data;
Step 2: Variable data = p.value; // < Custid, Tid, Item >
Step 3: For loop each data in p do
Step 4: Output < data.Item, (data.Custid, data.Tid) >;
Step 5: End for loop

Pseudo Code for Scanning Phase (Reducer)

Mapper Input <k, v>

sup, threshold
Step 1: Variable i = k; // storing input
Step 2: Variable sup = s; // used to store support
Step 3: For loop each value in v do
Step 4: Sup = find the support count of each item (i)
Step 5: End for loop
Step 6: For loop each < i, s > in sup do
Step 7: If s(i).support >= sup then
Step 8: Variable bit = create the bitmap of each item (i)
Step 9: Put < i, (data, bit) > to DHT
Step 10: End if
Step 11: End for loop

2. Mining Phase

In the mining phase the mapper task creates a lexical sequence tree for finding
patterns and to improve efficiency of limited depth dfs which is run on the tree
and the reducer task finds the support value for the patterns and finds the useful
patterns. These phases are further elaborated. In the scanning phase the data is
loaded on to the machine for scanning and mining. To avoid loss of data during
loaded on to the machine for scanning and mining. To avoid loss of data during
loading each mapper task and reducer task each of these tasks reads pre-
partitioned part of the data based on a memory chunk or no of lines. The mapper
task converts the data Tuples to key value pairs for faster processing. The tuples
are grouped based on itemID and the transaction IDs are stored in the value part.
For example, item,(tid,tid) The reducer task counts the support value for the
items and patterns which can be used to remove infrequent patterns. For
handling data between mappers and reducers the mapper and reducer with
identical keys handle the same set of data. A threshold would be identified to
remove the infrequent patterns. To remove infrequent patterns any pattern
having confidence value less than the threshold would be removed and the
reducer would store the data in a distributed hash table or a distributed file
system would be accessible to the next MapReduce job in the mining phase.

Pseudo Code for Mining Phase (Mapper)

Input: X, depth (d), maximum depth of subtree

Step 1: Variable LST; // Lexical Sequence Tree on loc machine
Step 2: LST = generate a subtree with root node is X and depth d
Step 3: For loop each node in LST do
Step 4: Variable bit-A_r = perform bit-AND operation
Step 5: If (bit-A_r > 0)then
Step 6: Output < node.pattern, (Cid, bit-A_r) >;
Step 7: End if
Step 8: End for

Pseudo Code for Mining Phase (Reducer)

Mining Input: a set of < pattern, (Custid, bit-A_r) > pairs

min_sup, minimum support threshold
Step 1: For loop each pattern do
Step 2: Freq = store net support count of this pattern;
Step 3: If freq >= min_sup then
Step 4: Output pattern < pattern, freq >;
Step 5: If this pattern is a leaf node in the current sub-extension-tree then
Step 6: Output pattern to the extensible set;
Step 7: End if loop
Step 8: End if loop
Step 9: End for loop
Here the main aim is to find all possible patterns in the dataset and to prune
away non-sequential and non-useful patterns to filter out the important
information from the dataset. We try to achieve this using a lexical sequence tree
which can generate all possible patterns existing in the dataset and then this tree
can be pruned away to find out useful or sequential patterns. The task of the
reducer phase is to merge the outputs from the mapper phase and reducer phase
prunes away the patterns with low support count. The reducer phase extracts the
depths and support count from the BitANDresult to gain info from the
preprocessed data form the mapper phase a reducer phase works on the output of
a unique mapper in parallel on a distributed environment constructed on mapper
processes running in parallel. We can conclude from this behavior that larger
datasets can be handled better by synchronizing master and slave nodes. The
better option is to always trigger the master node with recursive slave node
callouts.

Various parameters have been considered for performance evaluation in Pattern

Mining on various datasets. Some of them are:

• Memory scalability
• Work partitioning
• Load balancing.

Figure 2. Architecture diagram

Evaluation of performance of Iterative-MapReduce on diabetes dataset using
Hadoop with 1 master node and 2 slave nodes gave positive results. It took
reduced memory rounds and computed the patterns very efficiently. On
comparing with the existing methods this is much more scalable and accurate.
Generic implementation of i-MapReduce on Hadoop 1.0.3 and jdk-7u3 in a
virtual cloud environment consisting of 3 clusters gave a memory round 225ms.
First cluster serves as both master and slave node and the remaining 2 clusters
serve as slave nodes. All these evaluations which are carried out on machines
with 2.64GHz Intel Xeon CPU, 8GB main memory, and 1GB network
bandwidth produce similar outputs. It involves the process of feeding the (key,
value) pair to the mapper of the master node and then the slave node keeps on
repeating the iterative process till we achieve the confidence of 85% and
accuracy greater than 80%. With each iterative process the slave node uses the
output of first tool runner class and feeds as input for the next mapper task. On
investigating the results, it can be clearly seen that larger the number of Map
Reduce rounds, the better the Sequence tree of Mapper and thus this will help in
reducing the initialization costs and making enhanced and smooth dataflow.
Results have also shown that generally finding larger patterns will take more
time, so Iterative MapReduce focuses on finding small patterns and remembers
time, so Iterative MapReduce focuses on finding small patterns and remembers
them so that they can be further be used to propagate on other larger patterns and
this process in continued until we get the threshold accuracy and confidence
value specified by the user thus eliminating all unnecessary patterns.

Figure 3. Metric (Time, Space) comparison

Figure 4. Diabetic dependency on genes

FUTURE RESEARCH DIRECTIONS
For future work, further investigation needs to be done on improving the
computation time of the MapReduce algorithms. Clusters can be setup on
parallel GPUs and then observations can be made as to how the algorithm reacts
to a highly parallel environment with real time influx of data and better memory
management. Also intrinsic pattern pruning is one area which has a great scope
of future research. The cluster nodes need to be made smart enough to
automatically prune the uninteresting patterns from the dataset.

CONCLUSION
In this chapter we have discussed an optimized version of a traditional pattern
mining algorithm on a distributed setup using Hadoop framework and Iterative
MapReduce algorithms. A stepwise system to find and filter relevant patterns
from the data set using Iterative MapReduce is the major point discussion in the
chapter. The capacity to handle large data sets varies depending upon the
server’s capacity as well as the amount of commodity servers available. By the
usage of commodity server’s the data processing has been scaled out. Thus we
can say that MapReduce algorithms play a very pivotal role in pattern mining.
The distributed environment of the node clusters helps in smooth flow of data
accompanied by efficient computation and optimized resource utilization.

REFERENCES

Dean, J., & Ghemawat, S. (2008). MapReduce: Simplified data

processing on large clusters . Communications of the ACM , 51(1),
107–113. doi:10.1145/1327452.1327492

Demirkan, H., & Delen, D. (2013). Leveraging the capabilities of

service-oriented decision support systems: Putting analytics and big
data in cloud . Decision Support Systems , 55(1), 412–421.
doi:10.1016/j.dss.2012.05.048

Kumar, S. (2004, January 01). MEGA3: Integrated software for

Molecular Evolutionary Genetics Analysis and sequence alignment.
Briefings in Bioinformatics , 5(2), 150–163. doi:10.1093/bib/5.2.150

Labrinidis, A., & Jagadish, H. V. (2012). Challenges and

opportunities with big data . Proceedings of the VLDB Endowment
International Conference on Very Large Data Bases , 5(12), 2032–
2033. doi:10.14778/2367502.2367572

Reges, S., & Stepp, M. (2014). Building Java Programs . Pearson.

Ristoski, P., Mencía, E. L., & Paulheim, H. (2014, May). A hybrid

multi-strategy recommender system using linked open data. In
Semantic Web Evaluation Challenge (pp. 150-156). Springer
International Publishing. doi:10.1007/978-3-319-12024-9_19

Sankaranarayanan, S., & Perumal, T. P. (2014). Diabetic Prognosis

through Data Mining Methods and Techniques,” Intelligent
Computing Applications (ICICA), 2014 International Conference on,
162-166.

Taylor, R. C. (2010). An overview of the Hadoop/MapReduce/HBase

framework and its current applications in bioinformatics . BMC
Bioinformatics , 11.

Tukey, J. W. (1977). Exploratory data analysis. Academic Press.

White. (2012). Hadoop the definitive guide . O'Reilly Media, Inc.

 CHAPTER 8

Big Data for Satellite Image Processing:

Analytics, Tools, Modeling, and Challenges
Remya S.
VIT University, India

Ramasubbareddy Somula
VIT University, India

Sravani Nalluri
VIT University, India

Vaishali R.
VIT University, India

Sasikala R.
VIT University, India

ABSTRACT
This chapter presents an introduction to the basics in big data including
architecture, modeling, and the tools used. Big data is a term that is used for
serving the high volume of data that can be used as an alternative to RDBMS
and the other analytical technologies such as OLAP. For every application there
exist databases that contain the essential information. But the sizes of the
databases vary in different applications and we need to store, extract, and modify
these databases. In order to make it useful, we have to deal with it efficiently.
This is the place that big data plays an important role. Big data exceeds the
processing and the overall capacity of other traditional databases. In this chapter,
the basic architecture, tools, modeling, and challenges are presented in each
section.
section.

1. INTRODUCTION
Day by day, we see the data is rapidly increasing in many forms. We have some
traditional data processing software to process small quantity of data. But as
trillions of bytes of information is being processed per second, the traditional
software techniques fail in processing this data. We need to re-think of a solution
which can process this data. Now Big Data gives us a solution. Big Data is a
term used for creating, capturing, communicating, aggregating, storing and
analyzing large amounts of data. Many attempts encountered to quantify the
growth rate in the volume of data is called as Information Explosion.

Major milestones took place in the history of sizing data volumes plus the
evolution of the term Big Data. The following are some of them:

• In 1971, Arthur Miller stated in “The Assault on Privacy” that:

Too many information handlers seem to measure a man by the

number of bits of storage capacity his dossier will occupy.

• In April 1980, I.A.Tjomsland gave a talk titled “Where Do We Go From

Here?” at “Fourth IEEE Symposium on Mass Storage Systems” in which he
says:

Data expands to fill the space available, I believe that large

amounts of data are being retained because users have no
way of identifying obsolete data, the penalties for storing
obsolete data are less apparent than are the penalties for
discarding potentially useful data.

• In 1997, Michael Lesk publishes “How much information is there in this

world?” in which he concludes that:

There may be a few thousand petabytes of information all

told, and the production of tape and disk will reach that level
by the year 2000. So in only a few years, (a) we will be able to
save everything-no information will have to be thrown out,
and (b) the typical piece of information will never be looked
at by a human being.
(https://www.forbes.com/sites/gilpress/2013/05/09/a-very-
 (https://www.forbes.com/sites/gilpress/2013/05/09/a-very-
short-history-of-big-data/2/#1c3097c24343).

The term Big Data was coined in 1998 by Mr. John Mashey, Chief Scientist at
SGI. Even though Michael Cox and David Ellsworth seem to have used the term
‘Big Data’ in print, Mr. Mashey supposedly used the term in his various
speeches and that’s why he is crediting from coming up with Big Data. But some
various sources say that the first use of the term Big Data was done in an
academic paper-Visually Exploring Gigabyte Datasets in Realtime(ACM)
(OECD, 2015; Mark A. Beyer & Douglas Laney, 2012).

The following are the differentiators of Big Data over Traditional Business
Intelligence solutions:

• Data is retained in a distributed file system instead of on a central server.

• The processing functions are taken to the data rather than data being taken
to the functions.
• Data is of different formats, both structured as well as unstructured.
• Data is both realtime as well as offline data.
• Technology relies on massively parallel processing(MPP) concepts.

The Big Data challenges include capturing data, data storage, data analysis,
search, sharing, transfer, visualization, querying, updating and information
policy. Organizations have to compromise and balance against the
confidentiality requirements of the data. Organizations must determine how long
the data has to be retained. With the advent of new tools and technologies to
build big data solutions, availability of skills is a big challenge for CIO’s. A
higher level of proficiency in the data science is required to implement big data
solutions today because the tools are not user-friendly yet. (Bill Franks, 2012).

Analogous to the Cloud Computing architecture, the Big Data landscape can be
divided into four layers.

• Infrastructure as a Service (IaaS): This includes the storage servers, and

network as the base, inexpensive commodities of the big data stack.
• Platform as a Service (PaaS): The unstructured data stores and
distributed caches that can be logically queried using query languages
serves the platform layer of Big Data.
• Data as a Service (DaaS): The tools required to integrate the PaaS layer
using search engines, integration adapters, batch programs and so on is
 housed in this layer.
• Big Data Business Functions as a Service (BFaaS): Industries like
health, retail, e-commerce, energy and banking build applications that serve
a specific purpose and hold the DaaS layer for cross-cutting data functions.

Recent Gartner’s “Hype Cycle for Emerging Technologies”, visualizes the

absence of Big Data. Big Data got a permanent place in the Gartner Cycle for the
past 5 years in the emerging technologies, but shockingly it was removed in July
2016. The reason was given by Burton, Gartner Analyst that I would not
consider big data to be an emerging technology. This hype curve is very much
focused. I look at emerging trends.
(http://www.gartner.com/technology/research/hype-cycles/).

Here are some of the big data providers that are offering solutions in the specific
industries:

• The Securities Exchange Commission (SEC) is using big data to monitor

financial market activity. They are currently using network analytics and
NLP to catch illegal trade activity in financial markets.
• In Communications, Media and Entertainment, industry simultaneously
analyze customer data along with behavioral data to create detailed
customer profiles.
• Some hospitals are using Big Data techniques to use evidence-based
medicine as opposed to administering several medical/lab tests to all
patients who go to the hospital.
• Big Data has also been used in solving today’s manufacturing challenges
and to gain competitive advantage among other benefits.
• In public services, Big Data has a very wide range of applications
including energy exploitation, financial market analysis, fraud detection,
health related search and environmental protection.
• In the field of Insurance, Big Data is used to provide customer insights for
transparent and simpler products, by analyzing and predicting customer
behavior through data derived from social media, GPS-enabled devices and
CCTV footage.
• Smart meter readers allow data to be collected almost every 15 minutes as
opposed to once a day with the old meter readers. This granular data is
being used to analyze consumption of utilities better which allows for
improved customer feedback and better control of utilities use (David R.
Hardoon & Galit Shmueli, 2013).
We have 4 V’s of Big Data:

• Velocity: The data is increasing at a very fast rate. It is estimated that the
volume of data doubles every year.
• Variety: Now a days data is not stored in rows and columns i.e.,
structured format. We see data is being stored in the form of log files i.e.
unstructured.
• Volume: The amount of data which we deal with is of very large size of
peta bytes.
• Veracity: Explains the reliability of data (Foster Provost & Tom Fawcett,
2013).

2. BIGDATA MODELING (4 V’S)

Bigdata can be defined as a collection of large amount of data sets. Using
traditional tools it is very difficult to handle such a vast amount of datasets. This
is the reason that bigdata become important in storing and analyzing large
amount of data. These data can be mined by using data mining tools or DBMS
tools. Based on this the main components of big data can be termed as set of 4
V’s (P. Hitzler & K. Janowicz, 2013).

Figure 1. Modeling of Bigdata

2.1 Volume, Velocity, Variety, Veracity

• Volume refers to the amount of data

• Velocity refers to the rate at which data can be generated and transmitted
• Variety refers to the different types of data such as structured, semi
structured and unstructured
• Veracity refers to the integrity of the data

The bigdata differentiates from the traditional data based on these 4 components:

2.1.1 Volume

We live in the data age and in 2013 it was analyzed that the total volume of data
storage is 4.4 zeta bytes and in 2020 it will become 44 zeta bytes (1 zeta byte is
10 21 Bytes. These very much amount of data can need to be stored and
analyzed. These data are from different sources and also need to be combined for
better results.

2.1.2. Velocity

The data is from multiple sources and these sources are to be run parallel. Hence
the important next issue is how to run these sources parallel with very high speed
for data generation and transmission. For example consider a weather sensor
which collects the weather data from multiple sources in each and every hour.
These data is need to be move on to a particular storage and this data log is very
high. Traditional systems are not capable of doing this storage and frequent
movements.

2.1.3. Variety

The data sources can be of different types which can be collected from weather
sensors, social networks, stock exchange and smart phones. The data includes
text, images, audio, video or any other data logs. The data can be classified
mainly into three such as structured data, semi structured data and unstructured
data. Traditional distributed systems such as RDBMS, volunteer computing and
grid computing can handle only structured data. Bigdata differs from these
systems by it can handle semi structured and unstructured data also.

2.1.4. Veracity

The bigdata can handle huge amount of data and this data need to be correct
also. Hence the Veracity refers to how we can clean the data for data
preprocessing stage. The data need to be relevant and valuable (X. L. Dong & D.
Srivastava, 2013).

3. ARCHITECTURE OF BIGDATA
Bigdata is treated as set of tools for developing and analyzing scalable, reliable
and portable data. It serves as key for design infrastructure and solutions. It
interconnects the existing and organizing the resources and consist different
layers such as:

• Bigdata Sources: The location which produces the data.

 • Messaging and Storage: The facilities where the data is stored.
• Bigdata Analysis: The different tools for analyzing the different types of
data.

A bigdata architecture is designed so that it can handle processing, storage and

analysis of complex large data .Bigdata can handle processing of big data
sources, real time processing of data, predictive analytics using machine learning
approaches and exploration of interactive data (K.Bakshi, 2012).

The architecture includes the following components:

• Data Sources: Real time data sources, application of data sources such as
relational databases. Static files produced by webservers and other
applications.
• Data Storage: Bigdata is not the first distributed processing systems. But
it can store high volumes of data known as Data Lake, than other traditional
systems. Bigdata can prepare the data for analysis and then these analytical
data store can be used to serve different types of queries. The analytical
data store can also provide metadata abstraction and low latency NoSQL
technologies.
• Processing of Data: Bigdata provides interactive and batch processing
including real time applications. Bigdata solutions process the data files
using batch systems to filter and aggregate. If the solutions include real
time sources, the bigdata architect can include stream processing also. The
processed stream data is then written into an output sink (B. Ramesh,
2015).
• Service Orchestration: Most of the bigdata solutions consists of repeated
data processing operations, then transform the encapsulated source data.
The movement of data between multiple sources and destination and load
the processing data in to an analytical store is an important issue in
traditional systems. This can be automated by using service orchestration in
bigdata by using the tools such as sqoop and Oozie.

Figure 2. Architecture of Bigdata

4. TOOLS USED IN BIGDATA ANALYTICS

4.1 Big Data Analytics with GIS Datasets

A number of open source tools, frameworks and query languages have been
introduced to analyse big data. MongoDB is a famous NoSQL based data
analytics tool that provides option to visualize, analyse and explore datasets. In
this section let us explore a GIS dataset in Mongo Compass.

Here is the step by step implementation of the work:

Step 1: Install MongoDB Compass from mongodb.com as per the

instructions provided in the website.
Step 2: Configure the MongoDB compass with the following details to get
connected to the host.

Figure 3. Configuration screen of the mongoDB Compass

 Step 3: Add the 100YWeather dataset to the MongoDB Compass. The
dataset appears on the right side of the screen. Expand the title to see the
data collection. Click on the data collection to view the records inside.

Figure 4. MongoDB Compass

 Step 4: Document View. In mongoDB records are named as json
documents. The screen lists a huge collection of weather data in 250,000
json documents, with 5 Indexes and size of 403MB.

Figure 5. Document View

Schema View

Schema view lists the attributes information and visualizes the data with data
types.

Figure 6. Schema View

To get the view of the Schema click on the Schema Tab at and top and then click
on the green ‘Analyse’ button.

As shown in Figure 6, the query made to analyse the 100Y weather dataset has
returned 250,000 json documents

The different tools in bigdata are summarized in Table 1.

5. CHALLENGES IN BIGDATA
The term ‘Bigdata’ implies that it is a big volume of Data. As many industries
are generating data volumes from terabytes to petabytes, there developed a need
to process the raw data in to useful information. In order to provide benefits to
business and Information technology, the concern for the enhancement of big
data storage and processing architectures has increased. Potentially Bigdata
undergoes a lot of challenges with the characteristics, data analytics and
processing capabilities of the system. In this section, we mainly discuss about
the challenges in major research areas (X. Yi, F. Liu, J. Liu, & H. Jin, 2014).

5.1 Data Complexity

The Bigdata collects a large scale data from different sources and solve the
computational problems. The most significant characteristics of big data are
types of data patterns, different structures and complicated communication
between data samples. Big data refers to the data (or) Information which can
always be processed with advance technology like analytics, visualization
methods and can also find hidden pattern in order to make an accurate decision
rather than using additional compute system (“Global Data Center Traffic”).
Many Business organizations have large amount of data but that whole data is
used properly due to the lack of efficient systems so that the percentage of data
utilization is keep on decreasing (Oracle Big Data Strategy Guide). Now the
technology is growing day by day, the mobile devices and sensors playing an
important role for generating data and then stored. For example, people can
monitor working people away from office because the office people staying far
away from office (Cloudera's 100% Open Source Distribution of Hadoop). For
instance, railway companies decided to install sensors for every few feet in order
to monitor internal and external event which causes train to meet accidents. By
monitoring every event, railway people comes to know that which part is
required to replace and which one is repaired (James Manyika, Michael Chui,
Brad Brown, Jacques Bughin, Richard Dobbs, Charles Roxburgh, & Angela
Hung Byers, 2011). The roads are equipped with sensors to read the information
for reducing the chances of occurring natural disaster by analyzing and
predicting generated data (“Big Data, Big Impact: New Possibilities for
International Development”). Everyday few terabytes of data is generated from
every business organization, to deal with the traditional methods such as
information retrieval, discovery, analysis, sentimental analysis, but they are not
fit for huge data. Currently, we don’t have good knowledge on data distribution
law and association relation of big data. Also lack of deep understanding
association between data complexity and computational complexity arises. The
lack of the processing methods in big data and all the other aspects confine our
ability to implement new method and models to solve all these problems in big
data. The basic problem is to understand essential characteristics of complexity
of big data. Basic study on complexity theory will give clear view on complexity
and how they are formed, how they are associated with other pattern. By getting
clear understanding on complex theory, we can design and implement novel
models to resolve problems of big data complexity.

5.2 Computational Complexity

Big data includes three main features such as fast-changing, multiple sources,
huge volume . These features become difficult for traditional processing
methods such as machine learning, data retrieval, data analysis. New big data
computing tools are needed to overcome problems coming from independent and
identical distribution of data for generating accurate statistics. For addressing the
problems of big data, we require examining computational complexity and used
algorithms. The storage available for generating data through social websites is
not enough, because of the popularity of big data on storage and network.
Server’s offloading the resource intensive data into cloud and sending more data
into cloud for processing. Offloading data into cloud does not solve the problem.
But big data require getting all the data collected and retrieving from terabytes to
petabytes. Offloading the entire data into cloud will take large time and also data
is changing in every second which will make the data hard to be updated in real
time. Offloading data from the storage location to processing location can be
avoided by two ways: one is to process the storage location and second one is
offload only required data which will take more time to process. Building
indexes for storing data which will make easy the retrieving process and
moreover reducing processing time. In order to address computing complexity in
big data we need to understand about life cycle application of big data to study
of centralized processing mechanisms which depends upon the behavior of big
data. We need to get way from tradition approach of computing-centric to
distributed computing paradigms. We need to focus on new methods and data
analysis mechanisms for distributed streaming mechanisms. We are also
required to focus more on boot strapping and local computation, new algorithms
for handling large amount of data.
5.3 System Complexity

Big data can process high-volume of heterogonous data types and applications to
support research on big data. As big data processing are not enough to handle
generating high volume of data and real time requirements, these constraints
pose to establish new system architecture, processing system and energy
optimization model. Addressing the complexity problems will lead to designing
novel hardware and software frameworks for optimizing energy-consumption on
big data. Systems can process data that are all similar in size and structure, but
make difficult to process when data is presented in different patterns and sizes.
We need to conduct small scale research on all big data tools, different work
load conditions, different data types, various data pattern, and performance
evaluation in distributed environment, centralized environment, machine
learning algorithm for performance prediction, energy optimized algorithms,
energy consumption per unit and recursive work. We should focus on novel data
processing systems which is able to process all kinds of data in different
situations.

6. APPLICATIONS OF BIGDATA
• Big Data Analytics(BDA): This kind of application analyzes massive
data using parallel processing framework. BDA applications use sample
data in pseudo-cloud environment. After that they build in real cloud
environment with more processing power and large input data. These
applications utilize large data, which is unable to fit in hard drive. The data
is generated from different sources like traffic, social websites, the online
game information, and stock market and during international games.
• Clustering: User can easily identify group of people by using algorithms
such as k-means algorithms through points and click dialog and based on
specific data dimension. Clustering plays an important role in big data in
order to address group of people by considering customer type patient
documents, purchasing pattern, behavior products.
• Data Mining: The decision tree will help user to understand outcome and
relation between attributes in expected outcome. This decision tree reflects
the structure of that probability hidden in your data. Decision tree helps us
to predict fraud risk, online registrations, online shopping, and disease risk.
• Banking: In banking sector, the use of sensitive data leads to privacy
 issues. Research shows that more than 62% of bank employees are cautious
about their bank customer’s information to privacy issues. Distribution of
customer’s data to different branches also leads to security issues. The
investigation happened on banks data containing user’s sensitive
information such as earnings, savings, and insurance policies ended up in
the wrong hands. This discourages the customers in sharing personal details
in bank transactions.
• Stock: Data analytics can be used to detect fraud by establishing a
comprehensive system in data base during private stock exchange.
• Credit Cards: Credit card companies depend on in-data base analytics to
identify fraud transactions with accuracy and speed. This fraud transaction
deletion will follow up users sensitive data such as amount, location and
follow up before authenticate suspicious activity.

Enterprise

It will help the industry people exist around the world. Data doesn’t have to
move to work and back. It provides insight to business people to make accurate
decision for less expensive than traditional tools.

• Customer Goods: A manufacturer of customer products gather data

related to customer preferences and purchasing data of surveys, web tags,
customer call centers, the text taken up from online web sites, everything
that being said about product extract. By following this kind of analysis
business people will come to know that which product is to be succeeded
and others are to be failed. Finally, they can spot feature trends the products
in the marketing media.
• Agriculture: Using sensors, agriculture firm collect the data of efficiency
of crops. Initially, experts of agriculture field plant crops and run simulation
to find how crops react in different conditions. Research on agriculture will
collect data including various attributes, temperature, growth level, soil
composition, in order to find optimized environment for specific gene
types.
• Finance: The popular financial companies are focusing more on credit
scoring using third-party. Today financial institutions using their own
analysis for generating credit score for existing users using wide range of
information such as credit card, investments, transactions, savings.
• Economy: Hadoop can assist organization to perform low cost
transactions.
 • Telecom: People carry face reorganization technology in their pocket.
Android users use remember app to recognize (or) collect data related to
that snapped image from data base, while “I phone” users unblock their
devices with recognize me, this app developed widely to save nearly $2.5
million per year for solving forgotten –passwords.
• Health Care: Traditional health care industries has lagged behind than
the other industries in big data, every stack holder of hospital taking
decision independently rather than depending upon big data tools (James
Manyika, Michael Chui, Brad Brown, Jacques Bughin, Richard Dobbs,
Charles Roxburgh, & Angela Hung Byers, 2011).

Most health care stakeholders invested in information technology because of

accurate results. The traditional systems have limited ability to cause data
become more certain. The healthcare industries itself create issues. It becomes
difficult to share data among department even within hospital. The important
information of payee, pharmaceutical company will be available within single
department because organizations lack of knowledge on integrating data and
result obtained. Big data advances healthcare ability to work with enormous data
even though data is presented in different formats.

7. CONCLUSION
Bigdata is an important technology in our data era which can handle structured,
semi structured and unstructured data. It provides a viable solution for large and
complex data and become a challenge in nowadays. Each bigdata system
provides a massive power and for providing this different tools are used. This
paper presented the important aspects of big data, architecture and its
applications. However the challenges are also presented in this paper. It is clear
that now we are starting in the bigdata era and we have to discover many things
about big data for competing this data world.

REFERENCES
BakshiK. (2012). Considerations for big data: Architecture and
approach. Aerospace Conference, 1–7. 10.1109/AERO.2012.6187357

Big Data, Big Impact: New Possibilities for International

Development. (n.d.). World Economic Forum.

Bill Franks. (2012). Taming the big data tidal wave . Wiley.

Cloudera's 100% Open Source Distribution of Hadoop. (n.d.).

Retrieved from http://www.cloudera.com/content/clou
dera/en/products/cdh.html

David, R. (2013). Getting started with business analytics – insightful

decision making . Talor & Francis Group.

Dong, X. L., & Srivastava, D. (2013). Big data integration. In Data

Engineering (ICDE), 2013 IEEE 29th International Conference on,
1245–1248. 10.1109/ICDE.2013.6544914

Global data center traffic – Cisco Forecast Overview. (n.d.). Retrieved

from http://www.cisco.com/en/US/solutions/
collateral/ns341/ns525/ns537/ns705/ns
1175/Cloud_Index_White_Paper.html

Hitzler, P., & Janowicz, K. (2013). Linked Data, Big Data, and the 4th
Paradigm. Semantic Web , 4(3), 233–235.

Manyika, Chui, Brown, Bughin, Dobbs, Roxburgh, & Byers. (2011).

Big data: The next frontier for innovation, competition, and
productivity. McKinsey Global Institute.

Mark, A. (2012). Beyer and Douglas Laney. “The Importance of ‘Big

Data’: A Definition . Gartner.

OECD. (2015). Data-driven innovation: big data for growth and well-
being . Paris, France: OECD Publishing.

Oracle Big Data strategy guide. (n.d.). Retrieved from

http://www.oracle.com/us/technologies bigdatabigdata-strategy-
guide- 1536569.pdf

Provost & Fawcett. (2013). Data science for business. O’Reilly.

Ramesh, B. (2015). Big data architecture . In Big Data (pp. 29–59).

Springer. doi:10.1007/978-81-322-2494-5_2

Yi, X., Liu, F., Liu, J., & Jin, H. (2014). Building a network highway
for big data: Architecture and challenges . IEEE Network , 28(4), 5–
13. doi:10.1109/MNET.2014.6863125
 CHAPTER 9

Big Data Processing:

Application of Parallel Processing Technique to Big
Data by Using MapReduce
Abhishek Mukherjee
VIT University, India

Chetan Kumar
VIT University, India

Leonid Datta
VIT University, India

ABSTRACT
This chapter is a description of MapReduce, which serves as a programming
algorithm for distributed computing in a parallel manner on huge chunks of data
that can easily execute on commodity servers thus reducing the costs for server
maintenance and removal of requirement of having dedicated servers towards for
running these processes. This chapter is all about the various approaches towards
MapReduce programming model and how to use it in an efficient manner for
scalable text-based analysis in various domains like machine learning, data
analytics, and data science. Hence, it deals with various approaches of using
MapReduce in these fields and how to apply various techniques of MapReduce
in these fields effectively and fitting the MapReduce programming model into
any text mining application.

INTRODUCTION
In the steadily changing perception of information technology, data being
In the steadily changing perception of information technology, data being
gathered and looked through for business knowledge purposes have achieved
excessive levels. Welcome to the big data revolution. Big data is the term used
for data set so large or complex such that they cannot be modified or processed
in the conventional programming environment or softwares for generating
specific prediction or output thr, i.e., the day to day data processing application
softwares are not able to deal with them. This happens because the day to day
software systems have limitation of their process and also these huge chunks of
data contain so many outlier and exaggerated information that they are cleaned
and made fit for the processing before they are dealt with. For processing,
curing, storing, analyzing, decision making, transferring, visualizing, query
processing, updating using this huge hunk of data, the Big Data environment is
used.

Now we live in such a generation when huge chunks of data are generated at
every moment which is used by the business persons for analysis. If these data
are to be used for processing, then serialized way of processing will not give
result efficiently because that will consume time more than required. So, to
avoid this problem of the serialized processing, we go for parallel processing in
Big Data environment. This parallel processing helps us to reduce processing
time because the data sets are dealt in relatively small parts and they are merged
after that.

This is the brief idea about the chapter and we will see more details inside the
chapter.

BACKGROUND
While defining Big Data, it is very difficult to differentiate exactly between data
and Big data. The data can also be processed in this environment but that is not
suitable way for that. But big data handling is not at all suitable in the day to day
software system. In processing the Big data, the input data are having the
qualities like-Volume: The quantity of the data determines the weightage and
potential insight. Variety: The nature and variety of the data analyses how
eclectic the data is. Velocity: The data actually varies with time with variance
and this determines how useful the data is.Variability:Inconsistant data set of the
base analysis can hinder the handling.Veracity: The quality of captured data can
vary greatly, affecting accurate analysis. These qualities can be considered as
formal definition of the Big Data.
formal definition of the Big Data.

While dealing with the parallel processing, Parallel computing (Kumar, V.,
Grama, A., Gupta, A., & Karypis, G., 1994) is the specific type of computation
in any computing or processing environment in which many calculations or the
execution of processes are done simultaneously, i.e., large problems can be
divided into smaller problems for processing and then at the end, they are
merged for producing output. Parallel processing reduces the processing time
drastically because more than one portion of the data are being processed at the
same time.

MAIN FOCUS OF THE CHAPTER

The main focus of this chapter is to know about the modern day systems to keep
up and process all this information in a timely manner, new technologies have
been developed and old ones have been improved upon. Massively parallel
processing, or MPP (Metropolis, N., 1986) and MapReduce are such
technologies.

MPP manages the coordinated processing of huge datasets utilizing different

processors. This empowers fast rates of execution for many-sided inquiries
running against extensive information distribution centers. The primary reason
this technology came about is a result of the huge amounts of data that were
inundating applications not intended for such huge volumes. The need to process
this data for the reasons for investigation urged originators to create ultra-quick
preparing systems. Without the methods MPP employs, a query may take very
long time to finish, making present day business insight frameworks and data
warehouses not as much as helpful. MPP is at the heart of a wide range of sorts
of huge data solutions, and has made considerable advances as a critical
innovation. Amazon Redshift,is a fast, fully managed, petabyte-scale data
warehouse solution and a prevalent cloud-based information (Gupta, A.,
Agarwal, D., Tan, D., Kulesza, J., Pathak, R., Stefani, S., & Srinivasan, V.,
2015) warehousing arrangement, utilizes MPP design to accomplish to a great
degree quick inquiry execution. MPP is one of the five key execution
empowering agents of Redshift, alongside columnar data storage, data
compression, query optimization, and compiled code.

What MPP does is genuinely straightforward in principle. It separates large, hard

to manage datasets into effortlessly workable chunks, assigning each to a
processor. There could be upwards of hundreds or even thousands of processors
processor. There could be upwards of hundreds or even thousands of processors
dealing with chunks from the same dataset. When all the data has been
processed, results from the many different processors are combined for a final
result set. How MPP really accomplishes this is through an informing interface
the individual processors use to communicate with each other. Every processor
has its own working framework and memory. This permits everyone to work
autonomously, on its assigned section of the database, while loosely
communicating with alternate processors. It is hence that MPP frameworks are
known as a “shard nothing” framework. Since OS, memory, disk or whatever
else is not being shared among the individual nodes in a shared nothing design,
conceivable outcomes and points of interest for fine-tuning the framework can
be realized. The MPP framework can be scaled to the same number of extra
nodes as required, isolating the work further and accelerating the framework to
optimum levels. Additionally, since the processor of every node works
independently of the others, there is no bottleneck to hinder performance.

MPP is not by any means the only technology available to encourage the
handling of substantial volumes

of data. MapReduce, a part of the Apache Hadoop Venture, is another

innovation that accomplishes similar things MPP does, however with a few
contrasts. Actually, you may even say MPP and MapReduce are distant cousins.
At first glance, technological differences between the two may not appear to be
too far separated, but rather relying upon the requirements of your data
distribution center, picking one over the other could have a tremendous effect.

• Performance: The optimization and distribution components of MPP

permit it to deal with the distribution of data among the different nodes.
This rates up processing time considerably, making it a better performance
decision over MapReduce. Additionally, MPP databases fit in with the
ACID compliance which remains for atomicity (Coyle Jr, D. J., Chang, A.,
Malkemus, T. R., & Wilson, W. G., 1997) i.e. rollback of all updates is
required if the update of any node fails, consistency (Eswaran, K. P., Gray,
J. N., Lorie, R. A., & Traiger, I. L., 1976) i.e. the data satisfies certain
consistency constraints, isolation[14] i.e. isolation requires that transactions
observe the” latest” snapshot of the database, and durability (Elnikety, S.,
Dropsho, S., & Pedone, F., 2006) i.e. commits and making the effects of
transactions durable are performed in one single action. ACID ensures
database transactions are processed accurately and reliably. This is not
 something that is naturally implemented in Hadoop, giving MPP the
general execution edge over MapReduce.
• Scaling: The highly specialized hardware used by MPP frameworks make
adaptability a difficult and costly recommendation. MapReduce and
Hadoop, be that as it may, can be deployed to economical commodity
servers, permitting the groups of nodes to develop as required.
• Deployment and Maintenance: MPP is generally simple to deploy and
maintain. Hadoop and MapReduce, on the other hand, can end up being a
major implementation project requiring expensive and specialized expertise
that may not be available in-house.
• Data Restrictions: Unlike MPP, Hadoop and MapReduce can tackle
unstructured data without the need to process it before use. There is no need
to massage the data before it can be used. So, MapReduce is preferred
slightly more here.
• Language: The language behind MapReduce's control component is
basically Java. MPP uses SQL, making it easier to use and more cost
effective. SQL is an outstanding query language, for the most part used by
database experts, wiping out the need to procure expensive Hadoop experts.
Additionally, existing SQL-based business intelligence tools are supported
with MPP. This is not the situation with MapReduce, and alternative
solutions must be explored.

Is MPP better than MapReduce or vice versa? This depends on the goals of each
organization, for them they are different tools that suit different situations. As a
matter of fact, there are some organizations that use both MPP and MapReduce,
affording them the advantage of the best of both worlds.

ISSUES, CONTROVERSIES AND PROBLEM

• There is No Need to Distinguish Big Data Analytics From Traditional
Data Analytics.
• Hadoop is Not Always the Best Tool: There are situations when Hadoop
is not adequate for big data in organizations; while crunching real time data,
when the organizations have numerous data sources and business processes
that can’t be fit into a single data infrastructure and when the organizations
do not have the ability pool for complicated data science tools like
MapReduce.
• Size of the Data Does Not Matter as in Real Time Analytics, Data
 May be Changing. What is More Important is its Recency: Big data
may not be the right choice, while crunching real time data, because
Hadoop like tools process data in nodes. With a temporal measurements
varying over time, the quantity of data doesn’t matter unless it ties in with
time. What really matters is the sample size at a particular instance (which
is again a dimension) and not the overall sample size.
• Bigger Data are Not Always Better: Data adequacy plays a vital role
when we run samples across various dimensions. The quality of data used
for crunching makes the quality of insights. If the signal to noise proportion
is high, the exactness of results may shift for dirty samples
• Big Data Involves an Ethical Issue: Security. In the era of big data, the
debate between privacy and personalization will be progressing. Big Data is
a major ordeal today however privacy guidelines is similarly essential. It is
unethical if people are unaware that their data is analyzed.

Figure 1. Number of line vs processing time graph

Figure 2. 5 lac line processing

Figure 3. 10 lac line processing
SOLUTIONS AND RECOMMENDATIONS
All the problems stated above have various solutions one of them being map
reduce implementation of various analysis algorithms. But before this we need to
know about the phases of map reduce algorithm and how it works, the data flow
among the various phases and how it deals with processing the data parallel.
Map reduce is a software framework developed and used by Google™ (Schatz,
M. C., 2009) to support parallel processing tasks explicitly for intensive data
handling. It serves as a data warehouse to handle jobs parallel and permit large
scale distributed data analysis (Taylor, R. C., 2010). Each task in Map Reduce is
broken into the following main phases which are mapper, sort and shuffle and
finally reducer. A phase named combiner which can be used as a localized
reducer is also used between mapper and reducer for optimization purposes i.e.
to send lesser amount of data across the network. The output of the map tasks,
called the intermediate keys and values, are sent to the reducers. The processing
of data takes a collection of input key/value pairs, and produces a set of output
key/value pairs. As stated above the user decides the process flow according to
the map and reduce tasks where the map task basically aims to initially create the
key and value for later processing as required by the further stages. In the
intermediary sort and shuffle phases operate like a group by operator where all
the values of the same key are grouped together and thus finally it gives unique
keys with a tuple of values attached to each of it. The reducer phase finally does
the work of giving a key and value pair separated by a delimiter where the value
logic can be thought of as giving a single value from a tuple or list of values as
per the requirement of the program. As stated above many analysis tasks can be
done by the usage of map reduce some of which will be aggregation, data
filtering, data organization and finally a few machine learning algorithm (non-
iterative). This mainly deals with getting insights about the various statistical
parameters of data which are generally used to get various information about the
general nature of data variation, data summarization and range. A few
algorithms of each category are listed below

Mean: The arithmetic mean, is the measure of central location of data. In the
cases of a robust measure of central tendency will often provide a better estimate
of the mean of the data (Tukey, J. W., 1977). It can be measured by summing up
all the values of given set and dividing the sum by the number of values. Hence
we need to fit our calculations according to the map reduce algorithm so that we
can form tuples meant for mean calculation from a set of data. Here we deal with
numeric data only as summation of string values is not possible. Mean
calculation needs to assume a key for which all the values will be accumulated
and divided by number of terms. Here we will assume a dataset which contains
the values separated by ‘,’ containing the numbers separated by ‘,’ which aims at
finally finding the average value of each column. Hence we will assume two
phases initially(map, reduce):

1. Map Phase: As stated above the map phase reads the input file line by
line and hence the key and value pairs for the input of map phase serves as
the byte offset of each line as the key and the line as a whole as the value
associated with each key i.e. the byte offset. Usage of byte offset address
translation may be completed in as few as a single clock cycle (Senthil, G.,
2004). Hence this serves as a benchmark for iterating through the line and
performing required operations. Thus while writing the code we can assume
a variable which is retained in every call of the iterator function and hence
is can be programmed as a counter which is incremented as we keep getting
the values for any particular column. Hence map output will have many
 keys associated with many values corresponding to number of lines in the
input file. During the sort and shuffle phase all the values i.e. the numbers
in the same column are sorted according to the key and finally the output of
this phase is key associated with a tuple(list) of values which means that
each column number(key) is associated with all values corresponding to all
the values in their respective columns.
2. Reduce Phase: As stated above the input to the reduce phase is the
column number as a key and a list of values as the values associated to the
key. Hence now the reducer phase has access to all values in a particular
column as a list. Hence a running sum is to be maintained which retains its
value as long the iteration of the list of values takes place which keeps
adding the numbers and hence at the end of the iteration we have the final
sum with us in the variable. Thus the sum obtained in divided by the
number of values of the list and is finally associated to the input key i.e. the
column number. Hence the value of mean is obtained in this way for each
of the column.

This method may sound perfect but it has a huge flaw. Consider a huge dataset
with a large number of lines. Hence when the reducer is fed in with the list of
values each key will have as many values in its tuple set as the number of lines
in the data. This will increase the network traffic i.e. a large number of values
will need to serialized altogether at the end of sort and shuffle and will need to
be sent to the reduce phase. This can be optimized by the diving the file by the
basis of size/number of lines and by usage of the combiner phase which will
serve as a localized reducer here. To make it efficient storage of two things in
each value i.e. the average of each key and the number of terms for which the
average has been calculated. Thus instead of sending all the values associated to
the columns one can find the average of a subset of the columns and send the
average accompanied by the number of terms involved in calculating the sum.
Hence at the reducer end a list of values (depending on the number of pieces the
file was divided into) are obtained. Now to calculate the average each of the
value obtained for each key is multiplied with number of occurrences which is
present in the tuple. Summing up of all such values and division by the sum of
number of occurrences produces the mean of the data and hence optimization is
achieved.

KNN(k nearest neighbor): It is a very famous classification algorithm used to

specify particular classes for specific data available. KNN shows the maximum
accuracy as compared to the Naive Bayes and a few other classification based
algorithms (Bijalwan, V., Kumar, V., Kumari, P., &Pascual, J., 2014). T. Hence
algorithms (Bijalwan, V., Kumar, V., Kumari, P., &Pascual, J., 2014). T. Hence
since it classifies data it must use some metric for performing classification
operation. In this case we take Euclidian distance as the metric based for the
classification task. The basic algorithm follows the principle of taking the
parameter ‘k’ as input and finding all the classes of data which are sorted by
their distance from the concerned data point. After the classes of data are sorted
by distance the first k classes are taken in ascending order. Finally the
occurrence of each unique class is calculated. The concerned data point belongs
to the class which has maximum occurrence in these obtained ‘k’ nearest classes.
Since it is used as a non-iterative algorithm here we can frame phases according
to the map reduce framework for implementing knn algorithm.

1. Map Phase 1: In the first phase of the implementation which is a mapper

phase the data set is iterated to find the distance of each of the data present
in the input data set from the data value whose class variable is to be
calculated. Each of this distance has an associated class variable i.e. the
class to which the particular input data value from which the distance has
been calculated belongs to. The usage of a tree map comes handy here.
Implemented as a linked “binary tree” structure, very fast: O(log N) ; keys
are stored in sorted order (Reges, S., &Stepp, M., 2014).Tree map is
basically a modified form of a binary search tree which stores certain data
elements according to associated values which are generally numeric
values. This is because the values stored in binary search tree are stored by
comparing with the available values already stored previously. This data
structure is used so as to access and store each element in a sorted fashion
in an easier manner. Since the worst case complexity of finding and storing
any element in O(n) this a tree map is preferred. In this case the value
stored in the class variable and the key(based on which the class variable is
stored) associated with it is the distance associated to it. Since the value of
‘k’ is static for the process as soon as the size of the tree map exceeds ‘k’
the value with the largest distance is eliminated. This helps in storing only
the nearest ‘k’ class variables
2. Map Phase 2: The previous phase did the main task of storing the
nearest k class variables. This phase is very similar to the word count map
phase where each class variable is associated with ‘1’ as a value. Hence at
the end of this phase we have as key and value pairs with class variable
being the key and 1 as the value associated with the key
3. Reduce Phase 1: In this phase the sort shuffle outputs give key as each
unique class variable of the ‘k’ nearest data elements and ‘1’ as the tuple of
 values associated with these key elements where ‘1’ occurs as many times
as the occurrence of the class variable obtained from the map phase. In this
phase all the values associated with each key is added so finally the output
of this phase in each unique class variable as the key and the number of
times each class variable occurs in the ‘k’ nearest set.
4. Map Phase 3: This is a phase which gives the final output of the class to
which the data tuple belongs. Here again the tree map is used to store the
largest key i.e. the key occurring the most number of times.

Data Recommender System: A key building block for collaborative filtering

recommender systems is finding most occurring data in an input dataset
(Ristoski, P., Mencía, E. L., & Paulheim, H., 2014). This algorithm aims in
tagging documents which is basically based on the fact of how often any
particular word occurs in a document. This is especially useful for finding
keywords in a document and important parts of a document. Hence it gives
useful insights into data to find the important contents of input data. The initial
phase involves the word count algorithm. In this phase mainly the words in the
file as segregated from each other to carry out the frequency aggregation task for
the input file. In the next few phases the data is sorted on basis of the frequency
of each word as obtained. The detailed analysis and phase wise tasks are
explained as follows:

1. Map Phase 1: In this phase the data is read line by line and all the words
are collected in the forms of tuples where each words is linked with ‘1’
which in a way means the frequency of occurrence of that words in the file
and hence this phase mainly identifies all the words in the input file and
hence the main task of this phase is to identify all the words irrespective of
their frequency of occurrence in the input file and hence this constitutes the
first phase of the job.
2. Reduce Phase 1: This phase involves the output of the map shuffle and
sort phases. The shuffle and sort phases do a group by operation on the key
i.e. the words in the file which are obtained from the map phase. The sort
phase collected all keys(words) of same value together and hence by the use
of string comparison all the keys are sorted. In the shuffle phase all the
values of the same keys are collected together to form a tuple of values
(1’s) which effectively results in the number of 1’s associated with each key
as the number of times the word actually occurs in the file. Finally in the
reduce phase all the 1’s of each key are added together and hence the
frequency of each unique key is obtained by summing up of the values.
At the end of these two phases we have the frequency of each word occurring in
the input file and hence we get the output of this phase as the words was keys
and the frequency of occurrence of each word as the value. This thus makes the
keys as unique. Now for the final output the words i.e the keys need to be sorted
on the basis of their frequency obtained.

3. Map Phase 2: In the next steps the advantage of sort phase is taken
hence the values i.e. the frequency will now be used as a key and the word
which was the key earlier will now be used as a value. Hence the task of
this mapper is to exchange or swap the frequency as the key and the word
as the value
4. Reduce Phase 2: During the sort and shuffle phases the words with same
frequency are collected together in the form of tuples and hence the output
of sort and shuffle phases will be sorted frequency of words as key and all
the words with the same frequency together. Since reducer outputs single
key and value pairs instead of key and list of values which may occur in the
case of many words having the same frequency and. Hence when the
iteration across the list of values for the key i.e. the frequency occurs an
empty string is assumed initially which is concatenated as and when new
words are obtained from the list of values as obtained by the reducer. This
results in the formation of key as the frequency and the words which have
same frequency as the value. Special characters can be used as delimiters
between the unique words so that separation is obtained between the words
of same frequency are obtained. These phases can be optimized by the
usage of combiner phase between map phase 1 and reduce phase 1 so that
less amount of data is to transferred across the network hence optimizing
network traffic usage. The combiner serves as a localized reducer by
counting the frequency of all words in pieces of data which can later be
combined easily since it is commutative operation.

Ordered Sorting: The aim of this algorithm is to perform a optimized sorting of

the input dataset in a parallelized manner. So basically it aims at sorting data
based on a parallel key. Sorting is easy in sequential programming. Sequential
programming based sorting is easy but since map reduce uses a ‘divide and
conquer’ approach it becomes a bit difficult to make the process work in the map
reduce programming model. One approach may be to take all the input data and
sort data on particular parts of the data but that will not work as a global
approach since sorting is not a commutative and associative property. The type
of data flow is that partitions are required so that each partition is sorted locally
and when all these partitions as concatenated we get the data sorted as a whole
Sorted data has a number of useful properties. Sorted data provides many
insights into data by giving the range and timeline of data. Sorted data also helps
in performing search operations on data sets by giving binary search type
methods which is O(logn) complexity as compared to normal searching methods
which mostly have O(n) complexity. Hence the applications range from indexing
data and finding some data within specific ranges in a very efficient and
optimized manner. The main requirement of this is that the key should be
comparable type of data so that the sorting can be performed. This is performed
by the usage of keys which denote a specific section of data i.e. keys as so
constructed that data is evenly distributed among them and only data within
specific ranges occur within each partition. This will hence require a pre analysis
of data so as that the map reduce job gets to know about the data variation as a
whole and hence can construct keys for the partition operation. This will then
later help in assigning data to each of these partitions each of which can be
sorted locally and later the partitions can be sorted according to the constructed
key so that at the end we have the sorted data. The analyze phase may be
optional for some specific types of data like the datasets which have a unique
element in the data which can hence be used to find the exact range of a
particular parameter in data and hence the data can be partitioned on the basis of
that parameter very easily. The analyze phase if required samples data randomly
and hence the partitions are built over those samples. The mapper does a simple
random sampling. When the output is given only the key which is the sorted
value is given as output with the value associated with it as a null since the value
will not be associated to any other value. Only one reducer is used here since
that will come as a final sorted list and hence data handling will be easy in that
manner. Finally the practitioner phase will be executed for data range as
obtained. The mapper extracts the sort key in the same way as the analyze step.
However, this time the record itself is stored as the value instead of being
ignored. The custom practitioner takes in all values and sends to reducer based
on the ranges as decided by the analysis phase. The reducer just outputs the
values as obtained as the sorted values. The most important thing is that the
number of reducers will be equal to number of partitions which have been
decided by the analysis phase. This becomes a heavy operation because it
involves two phases one for finding the ranges of data and other for actually
sorting the data which is basically a more of a heavy phase since this actually
involves shuffling of the data as obtained. The actual cost of finding the
partitions will be very less since it will involve only one reducer and hence the
amortized will reduce by a great extent. The order step will involve a lot of data
travelling over the network and hence multiple reducers will make the job easier
and optimized.

Inverted Index: In recent years, large-scale image retrieval has been shown
remarkable potential in real-life applications (Nguyen, B. V., Pham, D., Ngo, T.
D., Le, D. D., & Duong, D. A., 2014). This type of job is basically used to list
important terms in a dataset and hence is useful for tagging important words in a
document so that specific values can be traced out. Most search engines carry
out this task for making efficient searches in the web. The mapper gives an
output of the desired unique keys attached with values. The partitioner phase is
used to determine where data is to be sent to a reducer after processing is done
and hence it makes the final processing task of the reducer in a more distributed
fashion. Hence finally the reducer receives a collection of distinct row identifiers
to link them back to the input keys. Finally the reducers can used with unique
delimiters to carry out the operations of separating the key and value pairs. The
final output has the values associated with unique IDs from the input file and
hence they can be associated with data elements from the file. The performance
hence is based upon the unique types of index keys and the cardinality of the
keys i.e. the number of values each key is associated with.

CONCLUSION
The golden age of data has arrived. The capacity to handle large data sets
problems vary by the server capacity as well as the amount of commodity
servers available. By the usage of commodity server the data processing has
been scaled out. This was only possible using innovations like map reduce and
mpp on the software scale. This mainly involves avoiding system level
integrations for the processing tasks and hence making development easier.

REFERENCES
Bijalwan, V., Kumar, V., Kumari, P., & Pascual, J. (2014). KNN
based machine learning approach for text and document mining.
International Journal of Database Theory and Application , 7(1), 61–
70. doi:10.14257/ijdta.2014.7.1.06
Coyle, D. J., Jr., Chang, A., Malkemus, T. R., & Wilson, W. G.
(1997). U.S. Patent No. 5,630,124. Washington, DC: U.S. Patent and
Trademark Office.

Elnikety, S., Dropsho, S., & Pedone, F. (2006, April). Tashkent:

Uniting durability with transaction ordering for high-performance
scalable database replication. Operating Systems Review , 40(4), 117–
130. doi:10.1145/1218063.1217947

Elnikety, S., Pedone, F., & Zwaenepoel, W. (2005, October).

Database replication using generalized snapshot isolation. In Reliable
Distributed Systems, 2005. SRDS 2005. 24th IEEE Symposium on (pp.
73-84). IEEE. 10.1109/RELDIS.2005.14

Eswaran, K. P., Gray, J. N., Lorie, R. A., & Traiger, I. L. (1976). The
notions of consistency and predicate locks in a database system.
Communications of the ACM , 19(11), 624–633.
doi:10.1145/360363.360369

GuptaA.AgarwalD.TanD.KuleszaJ.PathakR.StefaniS.SrinivasanV.
(2015, May). Amazon Redshift and the case for simpler data
warehouses. In Proceedings of the 2015 ACM SIGMOD International
Conference on Management of Data (pp. 1917-1923). ACM.
10.1145/2723372.2742795

Kumar, V., Grama, A., Gupta, A., & Karypis, G. (1994). Introduction
to parallel computing: design and analysis of algorithms (Vol. 400).
Redwood City, CA: Benjamin/Cummings.

Metropolis, N. (1986). Massively parallel processing. Journal of

Scientific Computing , 1(2), 115–116. doi:10.1007/BF01061388

Nguyen, B. V., Pham, D., Ngo, T. D., Le, D. D., & Duong, D. A.
(2014, December). Integrating spatial information into inverted index
for large-scale image retrieval. In Multimedia (ISM), 2014 IEEE
International Symposium on (pp. 102-105). IEEE. doi:10.1007/978-3-
319-12024-9_19

Reges, S., & Stepp, M. (2014). Building Java Programs . Pearson.

Ristoski, P., Mencía, E. L., & Paulheim, H. (2014, May). A hybrid

multi-strategy recommender system using linked open data . In
Semantic Web Evaluation Challenge (pp. 150–156). Springer
International Publishing.

Schatz, M. C. (2009). CloudBurst: Highly sensitive read mapping

with MapReduce. Bioinformatics (Oxford, England) , 25(11), 1363–
1369. doi:10.1093/bioinformatics/btp236

Senthil, G. (2004). U.S. Patent No. 6,721,869. Washington, DC: U.S.

Patent and Trademark Office.

Taylor, R. C. (2010). An overview of the Hadoop/MapReduce/HBase

framework and its current applications in bioinformatics. BMC
Bioinformatics , 11.

Tukey, J. W. (1977). Exploratory data analysis. Academic Press.

 CHAPTER 10
Image Processing CSE-4019 Distinction Between Fake Note and a
Real Note

Shreya Tuli
VIT University, India

Gaurav Sharma
VIT University, India

Nayan Mishr
VIT University, India

ABSTRACT
Big data is a term for data sets that are so large or complex that traditional data
processing application software is inadequate to deal with them. Its challenges
include capturing data, data storage, data analysis, search, sharing, transfer,
visualization, querying, updating, and information privacy. Lately, the term big
data tends to refer to the use of predictive analytics, user behavior analytics, or
certain other advanced data analytics methods that extract value from data, and
seldom to a particular size of data set. In this chapter, the authors distinguish
between fake note and a real note and would like to take it to a level where it can
be used everywhere. Its data after the detection of the fakeness and the real note
can be stored in the database. The data to store will be huge. To overcome this
problem, we can go for big data. It will help to store large amounts of data in no
time. The difference between real note and fake note is that real note has its thin
strip to be more or less continuous while the fake strip has fragmented thin lines
in the strip. One could say that the fake note has more than one line in the thin
strip while the real note only has one line. Therefore, if we see just one line, it is
real, but if we see more than one line, it is fake. In this chapter, the authors use
foreign currency.

INTRODUCTION
There are 6 different steps in order to distinguish between fake and original
currency. In this project, we are using one of the famous segmentation
techniques named thresholding. And also, our project involves various process
like opening and closing etc…. In this we are using a real note and 2 fake notes
and are differentiating between them based on the black strips it has.

Thresholding

It is the simplest method of image segmentation. From a grayscale image,

thresholding can be used to create binary images. The simplest thresholding
methods replace each pixel in an image with a black pixel if the image intensity
is less than some fixed constant T, or a white pixel if the image intensity is
greater than that constant. In the example image on the right, this results in the
dark tree becoming completely black, and the white snow becoming completely
white.

Opening

In mathematical morphology, opening is the dilation of the erosion of a set A by

a structuring element B. Together with closing, this serves in computer vision
and image processing as a basic workhorse of morphological noise removal.
Opening removes small objects from the foreground (usually taken as the bright
pixels) of an image, placing them in the background, while closing removes
small holes in the foreground, changing small islands of background into
foreground. These techniques can also be used to find specific shapes in an
image. Opening can be used to find things into which a specific structuring
element can fit (edges, corners, ...).

Closing

In mathematical morphology, the closing of a set (binary image) A by a

structuring element B is the erosion of the dilation of that set, In image
processing, Closing is, together with opening, the basic workhorse of
morphological noise removal. Opening removes small objects, while closing
removes small holes. This process is to ensure that any disconnected regions that
are near larger regions get connected to each other.

Figure 1. Architecture
PROCEDURE
Let us now see the different steps involved:

Step 1: Read in the image.

In this step, we are going to read the particular image from a given location, and
we also resize the image if it is not in a desired size. Here we are going to
directly read our images from desktop using command “imread”.

The code for this

is: - clear all; close all;

Ireal = imread('C:\Users\Bunny\Desktop\SqbnIm.jpg');
% Real
Ifake = imread('C:\Users\Bunny\Desktop\2U3DEm.jpg');
% Fake
Ifake2 = imread('C:\Users\Bunny\Desktop\SVJrwaV.jpg');
% Fake #2
% //Resize so that we have the same dimensions as the
other images
Ifake2 = imresize (Ifake2, [160 320], 'bilinear'.

The output of this step is:

Here the first one is the real image and the rest two are fake ones (Figure 2).

Figure 2.
 Step 2: Extracting black strips.

In this step as we initially we have

Discussed that our project is going to work on the principle number of black
strips. If black strips are only 1 then it is original. And if we they are not equal to
1 then it is fake image.

The code for this is:

BlackStripReal = Ireal (:195:215,:);

blackStripFake = Ifake(:195:215,:);
blackStripFake2 = Ifake2(:,195:215,:);
Figure (1);
subplot(1,3,1);
imshow(blackStripReal);
title('Real');
subplot (1,3,2);
imshow(blackStripFake);
title('Fake');
subplot (1,3,3);
imshow(blackStripFake2);
title('Fake #2');
In this code, we use command “black strip” for the
detection of black strips in image. Here By looking at
all of the notes, the black strip value hover between
the 195th column to the 215th column.
This is if each image has 320 columns. And subplot
here denotes the 1*3 matrix grid and the particular
axes whether that can be 1,2,3 for real, fake,fake2.

The output of this step is: (Figure 3).

Figure 3.

Step 3: Converting RGB image into gray level and then thresholding

Here in this step we are going to convert our RGB image into gray level image
by using the command “rgb2gray”. And again, subplot means the same as
discussed above. And here we perform thresholding after image is converted into
gray scale image. It is done by first converting image into binary by using
command “im2bw” and using a threshold value of 30.

Here we used 30 heuristically as this was predominantly the intensity that was
for the black strip consisted of.

The code for this step is:

The code for this step is:

BlackStripReal =
rgb2gray(blackStripReal);
blackStripFake = rgb2gray(blackStripFake);
blackStripFake2 = rgb2gray(blackStripFake2);
Figure (2);
subplot (1,3,1);
imshow (blackStripReal);
title('Real');
subplot (1,3,2);
imshow(blackStripFake);
title('Fake');
subplot (1,3,3);
imshow(blackStripFake2);
title ('Fake #2');

The output for this is: (Figure 4).

Figure 4.

For thresholding: -

blackStripRealBW = im2bw (blackStripReal, 30/255);

blackStripFakeBW = im2bw (blackStripFake, 30/255);
blackStripFake2BW = ~im2bw (blackStripFake2, 30/255);
figure (3);
subplot (1,3,1);
imshow(blackStripRealBW);
title('Real');
subplot (1,3,2);
imshow(blackStripFakeBW);
title('Fake');
subplot (1,3,3);
imshow(blackStripFake2BW);
title ('Fake #2');
Here again we use subplot for the same reason as
discussed above:

The output for this: (Figure 5).

Figure 5.

Step 4: Opening.

In this step we are doing area opening of the image like by specifying a larger
area of about 100 to ensure that the image will get rid of any spurious noisy and
isolated pixels. You'll notice that for each of the images, there are some noisy
pixels on the edges. We perform this opening operation using the command
“bwareaopen”. This function removes pixel areas in a black and white image
that have less than a certain area.

The code for this is:

The code for this is:

areaopenReal = bwareaopen (blackStripRealBW, 100);

subplot(1,3,1);
imshow(areaopenReal);
title('Real');
subplot (1,3,2);
areaopenFake = bwareaopen (blackStripFakeBW, 100);
imshow(areaopenFake);
title('Fake');
subplot (1,3,3);
areaopenFake2 = bwareaopen (blackStripFake2BW, 100);
imshow(areaopenFake2);
title ('Fake #2');

Here also again we use subplot for the same reason discussed above.

The output for this is: (Figure 6).

Figure 6.

Step 5: Post Process.

This is nothing but closing process. Here in this process we use a square
structuring element of 5 x 5 to ensure that any disconnected regions that are near
larger regions get connected to each other. For this process to be done we use
command “imclose”.
command “imclose”.

The code for this process is:

se = strel ('square', 5);

BWImageCloseReal = imclose (areaopenReal, se);
BWImageCloseFake = imclose (areaopenFake, se);
BWImageCloseFake2 = imclose (areaopenFake2, se);
Figure(5);
Subplot (1,3,1);
imshow(BWImageCloseReal);
title('Real');
Subplot (1,3,2);
imshow(BWImageCloseFake);
title('Fake');
Subplot (1,3,3);
imshow(BWImageCloseFake2);
title ('Fake #2');

Here again we use subplot for the same reason as discussed above.

The output for this is: (Figure 7).

Figure 7.

Step 6: Counting the total number of objects in this strip.

The last step is to simply count the number of black lines in each image. If there
is just 1, this denotes that the bank note is real, while if there is more than 1, this
denotes that the bank note is fake. For this we use the command “bwlabel” and
use the second parameter to count how many objects are there.

The code for this is:

[~,countReal] = bwlabel(BWImageCloseReal);
[~,countFake] = bwlabel(BWImageCloseFake);
[~,countFake2] = bwlabel(BWImageCloseFake2);
disp (['The total number of black lines for the real
note is: ' num2str(countReal)]);
Disp (['The total number of black lines for the fake
note is: ' num2str(countFake)]);
Disp (['The total number of black lines for the second
fake note is: ' num2str(countFake2)]);
Finally, we use “disp” to display the output:
The total number of black lines for the real note is:
1
The total number of black lines for the fake note is:
2
The total number of black lines for the second fake
note is: 0
FULL MATLAB CODE
clear all;
close all;
Ireal = imread('C:\Users\Bunny\Desktop\SqbnIm.jpg'); %
Real
Ifake = imread('C:\Users\Bunny\Desktop\2U3DEm.jpg'); %
Fake
Ifake2 = imread('C:\Users\Bunny\Desktop\SVJrwaV.jpg');
% Fake #2
% //Resize so that we have the same dimensions as the
other images
Ifake2 = imresize (Ifake2, [160 320], 'bilinear');
%% //Extract the black strips for each image
blackStripReal = Ireal(:,195:215,:);
blackStripFake = Ifake(:,195:215,:);
blackStripFake2 = Ifake2(:,195:215,:);
Figure (1);
subplot (1,3,1);
imshow(blackStripReal);
title('Real');
subplot (1,3,2); imshow(blackStripFake);
title('Fake');
subplot (1,3,3); imshow(blackStripFake2); title ('Fake
#2');
%% //Convert into grayscale then threshold
blackStripReal = rgb2gray(blackStripReal);
blackStripFake = rgb2gray(blackStripFake);
blackStripFake2 = rgb2gray(blackStripFake2);
Figure (2);
subplot (1,3,1);
imshow (blackStripReal);
title ('Real');
subplot (1,3,2);
imshow (blackStripFake);
title ('Fake');
subplot (1,3,3);
imshow (blackStripFake2);
title ('Fake #2');
%% //Threshold using about intensity 30
blackStripRealBW = ~im2bw (blackStripReal, 30/255);
blackStripFakeBW = ~im2bw (blackStripFake, 30/255);
blackStripFake2BW = ~im2bw (blackStripFake2, 30/255);
figure (3);
subplot (1,3,1);
imshow (blackStripRealBW);
title ('Real');
subplot (1,3,2);
imshow (blackStripFakeBW);
title ('Fake');
subplot (1,3,3);
imshow (blackStripFake2BW);
title ('Fake #2');
%% //Area open the image
Figure (4);
areaopenReal = bwareaopen (blackStripRealBW, 100);
subplot (1,3,1);
imshow (areaopenReal);
title ('Real');
subplot (1,3,2);
areaopenFake = bwareaopen (blackStripFakeBW, 100);
imshow (areaopenFake);
title ('Fake');
subplot (1,3,3);
areaopenFake2 = bwareaopen (blackStripFake2BW, 100);
imshow (areaopenFake2);
title ('Fake #2');
%% //Post-process
se = strel ('square', 5);
BWImageCloseReal = imclose (areaopenReal, se);
BWImageCloseFake = imclose (areaopenFake, se);
BWImageCloseFake2 = imclose (areaopenFake2, se);
Figure (5);
Subplot (1,3,1);
Imshow (BWImageCloseReal);
Title ('Real');
Subplot (1,3,2);
Imshow (BWImageCloseFake);
Title ('Fake');
Subplot (1,3,3);
Imshow (BWImageCloseFake2);
Title ('Fake #2');
%% //Count the total number of objects in this strip
[~, countReal] = bwlabel (BWImageCloseReal);
[~, countFake] = bwlabel (BWImageCloseFake);
[~, countFake2] = bwlabel (BWImageCloseFake2);
Disp (['The total number of black lines for the real
note is: ' num2str(countReal)]);
Disp (['The total number of black lines for the fake
note is: 'num2str(countFake)]);
disp(['The total number of black lines for the second
fake note is: '
num2str(countFake2)]);
RESULTS

Figure 8.
As seen above these are the results that we obtain while we run this process.

CONCLUSION
The conclusion is that we are using various image processing techniques to
determine whether a note is fake or real. We are telling this based on number of
black strips that are present on the image. We have used the processes like
thresholding, opening, closing etc. Here, I would like to conclude the big data
helps us to overcome the problem of storing a large amount of data instantly and
with a great efficiency.

Big Data depends on the following 3V’s:

• Volume: Big data doesn't sample; it just observes and tracks what
happens.
• Velocity: Big data is often available in real-time.
• Variety: Big data draws from text, images, audio, video; plus it completes
missing pieces through a fusion.

FUTURE SCOPE
This project has a good future scope. Now-a-days most of the people print these
fake notes which are illegal. So, this process can be used by the bank people to
fake notes which are illegal. So, this process can be used by the bank people to
detect which is the real and fake currency which is very helpful.

Bid Data in the coming time will play a major role in storing a large amount of
data for storing the information of how many fake and how many real notes were
detected in the set.

Figure 9.

REFERENCES
Gunaratna, Kodikara, & Premaratne. (2008). ANN based currency
recognition system using compressed gray scale and application for
Sri Lankan currency notes-SLCRec. Proceedings of World Academy
of Science, Engineering and Technology, 35, 235-240.
Guo, Zhao, & Cai. (2010). A reliable method for paper currency
recognition based on LBP. In Network Infrastructure and Digital
Content, 2010 2nd IEEE International Conference on. IEEE.

Hassanpour, Yaseri, & Ardeshiri. (2007). Feature extraction for paper

currency recognition. In Signal Processing and Its Applications, 2007.
ISSPA 2007. 9th International Symposium on. IEEE.

Zhang, E.-H. (2003). Research on paper currency recognition by

neural networks. In Machine Learning and Cybernetics, 2003
International Conference on (vol. 4). IEEE.
Related References
To continue our tradition of advancing knowledge management and discovery
research, we have compiled a list of recommended IGI Global readings. These
references will provide additional information and guidance to further enrich
your knowledge and assist you with your own research and future publications.

Abril, R. M. (2011). The quality attribution in data, information and

knowledge . In Schwartz, D., & Te’eni, D. (Eds.), Encyclopedia of
knowledge management (2nd ed.; pp. 1343–1354). Hershey, PA: IGI
Global. doi:10.4018/978-1-59904-931-1.ch129

Abufardeh, S. (2013). KM and global software engineering (GSE) . In

Saeed, S., & Alsmadi, I. (Eds.), Knowledge-based processes in
software development (pp. 12–34). Hershey, PA: IGI Global.
doi:10.4018/978-1-4666-4229-4.ch002

Aggestam, L., Backlund, P., & Persson, A. (2010). Supporting

knowledge evaluation to increase quality in electronic knowledge
repositories. International Journal of Knowledge Management , 6(1),
23–43. doi:10.4018/jkm.2010103002

Aggestam, L., Backlund, P., & Persson, A. (2012). Supporting

knowledge evaluation to increase quality in electronic knowledge
repositories . In Jennex, M. (Ed.), Conceptual models and outcomes of
advancing knowledge management: New technologies (pp. 24–44).
Hershey, PA: IGI Global. doi:10.4018/978-1-4666-0035-5.ch002

Aiken, P., Gillenson, M., Zhang, X., & Rafner, D. (2011). Data
management and data administration: Assessing 25 years of practice.
Journal of Database Management , 22(3), 24–45.
doi:10.4018/jdm.2011070102

Akabawi, S., & Hodeeb, H. (2013). Implementing business

intelligence in the dynamic beverages sales and distribution
environment . In Khosrow-Pour, M. (Ed.), Cases on performance
measurement and productivity improvement: Technology integration
and maturity (pp. 194–221). Hershey, PA: IGI Global.
doi:10.4018/978-1-4666-2618-8.ch010

Al-Busaidi, K. A. (2011). A social and technical investigation of

knowledge utilization from a repository knowledge management
system . In Al-Shammari, M. (Ed.), Knowledge management in
emerging economies: Social, organizational and cultural
implementation (pp. 122–139). Hershey, PA: IGI Global.
doi:10.4018/978-1-61692-886-5.ch007

Al-Busaidi, K. A. (2012). The impact of supporting organizational

knowledge management through a corporate portal on employees and
business processes . In Jennex, M. (Ed.), Conceptual models and
outcomes of advancing knowledge management: New technologies
(pp. 208–229). Hershey, PA: IGI Global. doi:10.4018/978-1-4666-
0035-5.ch011

Alaraifi, A., Molla, A., & Deng, H. (2013). An empirical analysis of

antecedents to the assimilation of sensor information systems in data
centers. International Journal of Information Technologies and
Systems Approach , 6(1), 57–77. doi:10.4018/jitsa.2013010104

Alguezaui, S., & Filieri, R. (2010). Social capital: Knowledge and

technological innovation . In López Sáez, P., Castro, G., Navas López,
J., & Delgado Verde, M. (Eds.), Intellectual capital and technological
innovation: Knowledge-based theory and practice (pp. 271–296).
Hershey, PA: IGI Global. doi:10.4018/978-1-61520-875-3.ch013
Alhashem, A., & Shaqrah, A. A. (2012). Exploring the relationship
between organizational memory and business innovation.
International Journal of Knowledge-Based Organizations , 2(3), 32–
46. doi:10.4018/ijkbo.2012070102

Allan, M. B., Korolis, A. A., & Griffith, T. L. (2011). Reaching for

the moon: Expanding transactive memory’s reach with wikis and
tagging . In Jennex, M. (Ed.), Global aspects and cultural perspectives
on knowledge management: Emerging dimensions (pp. 144–156).
Hershey, PA: IGI Global. doi:10.4018/978-1-60960-555-1.ch010

Alsmadi, I., & Alda, S. (2013). Knowledge management and semantic

web services . In Saeed, S., & Alsmadi, I. (Eds.), Knowledge-based
processes in software development (pp. 35–48). Hershey, PA: IGI
Global. doi:10.4018/978-1-4666-4229-4.ch003

Alstete, J. W., & Meyer, J. P. (2011). Expanding the model of

competitive business strategy for knowledge-based organizations.
International Journal of Knowledge-Based Organizations , 1(4), 16–
31. doi:10.4018/ijkbo.2011100102

Alstete, J. W., & Meyer, J. P. (2013). Expanding the model of

competitive business strategy for knowledge-based organizations . In
Wang, J. (Ed.), Intelligence methods and systems advancements for
knowledge-based business (pp. 132–148). Hershey, PA: IGI Global.
doi:10.4018/978-1-4666-1873-2.ch008

Alves da Silva, N. S., Alvarez, I. M., & Rogerson, S. (2011).

Glocality, diversity and ethics of distributed knowledge in higher
education . In Morais da Costa, G. (Ed.), Ethical issues and social
dilemmas in knowledge management: Organizational innovation (pp.
131–159). Hershey, PA: IGI Global. doi:10.4018/978-1-61520-873-
9.ch009
Anantatmula, V. S., & Kanungo, S. (2011). Strategies for successful
implementation of KM in a university setting . In Jennex, M., &
Smolnik, S. (Eds.), Strategies for knowledge management success:
Exploring organizational efficacy (pp. 262–276). Hershey, PA: IGI
Global. doi:10.4018/978-1-60566-709-6.ch014

Andreu, R., & Sieber, S. (2011). External and internal knowledge in

organizations . In Schwartz, D., & Te’eni, D. (Eds.), Encyclopedia of
knowledge management (2nd ed.; pp. 298–307). Hershey, PA: IGI
Global. doi:10.4018/978-1-59904-931-1.ch029

Andriessen, D. (2011). Metaphor use in knowledge management . In

Schwartz, D., & Te’eni, D. (Eds.), Encyclopedia of knowledge
management (2nd ed.; pp. 1118–1124). Hershey, PA: IGI Global.
doi:10.4018/978-1-59904-931-1.ch107

Angelopoulos, S., Kitsios, F., & Moustakis, V. (2012).

Transformation of management in the public sector: Exploring the
strategic frameworks of e-government . In Papadopoulos, T., &
Kanellis, P. (Eds.), Public sector reform using information
technologies: Transforming policy into practice (pp. 44–58). Hershey,
PA: IGI Global. doi:10.4018/978-1-60960-839-2.ch003

Anselma, L., Bottrighi, A., Molino, G., Montani, S., Terenziani, P., &
Torchio, M. (2013). Supporting knowledge-based decision making in
the medical context: The GLARE approach . In Wang, J. (Ed.),
Intelligence methods and systems advancements for knowledge-based
business (pp. 24–42). Hershey, PA: IGI Global. doi:10.4018/978-1-
4666-1873-2.ch002

Arh, T., Dimovski, V., & Blažic, B. J. (2011). ICT and web 2.0
technologies as a determinant of business performance . In Al-
Mutairi, M., & Mohammed, L. (Eds.), Cases on ICT utilization,
practice and solutions: Tools for managing day-to-day issues (pp. 59–
77). Hershey, PA: IGI Global. doi:10.4018/978-1-60960-015-0.ch005

Ariely, G. (2011). Operational knowledge management in the military

. In Schwartz, D., & Te’eni, D. (Eds.), Encyclopedia of knowledge
management (2nd ed.; pp. 1250–1260). Hershey, PA: IGI Global.
doi:10.4018/978-1-59904-931-1.ch119

Assefa, T., Garfield, M., & Meshesha, M. (2014). Enabling factors for
knowledge sharing among employees in the workplace . In Al-
Bastaki, Y., & Shajera, A. (Eds.), Building a competitive public sector
with knowledge management strategy (pp. 246–271). Hershey, PA:
IGI Global. doi:10.4018/978-1-4666-4434-2.ch011

Assudani, R. H. (2011). Negotiating knowledge gaps in dispersed

knowledge work. International Journal of Knowledge-Based
Organizations , 1(3), 1–21. doi:10.4018/ijkbo.2011070101

Assudani, R. H. (2013). Negotiating knowledge gaps in dispersed

knowledge work . In Wang, J. (Ed.), Intelligence methods and
systems advancements for knowledge-based business (pp. 75–96).
Hershey, PA: IGI Global. doi:10.4018/978-1-4666-1873-2.ch005

Atkins, R. (2011). Supply chain knowledge integration in emerging

economies . In Al-Shammari, M. (Ed.), Knowledge management in
emerging economies: Social, organizational and cultural
implementation (pp. 104–121). Hershey, PA: IGI Global.
doi:10.4018/978-1-61692-886-5.ch006

Aung, Z., & Nyunt, K. K. (2014). Constructive knowledge

management model and information retrieval methods for software
engineering . In Software design and development: Concepts,
methodologies, tools, and applications (pp. 253–269). Hershey, PA:
IGI Global. doi:10.4018/978-1-4666-4301-7.ch014
Aziz, M. W., Mohamad, R., & Jawawi, D. N. (2013). Ontology-based
service description, discovery, and matching in distributed embedded
real-time systems . In Nazir Ahmad, M., Colomb, R., & Abdullah, M.
(Eds.), Ontology-based applications for enterprise systems and
knowledge management (pp. 178–190). Hershey, PA: IGI Global.
doi:10.4018/978-1-4666-1993-7.ch010

Badia, A. (2011). Knowledge management and intelligence work: A

promising combination . In Schwartz, D., & Te’eni, D. (Eds.),
Encyclopedia of knowledge management (2nd ed.; pp. 612–623).
Hershey, PA: IGI Global. doi:10.4018/978-1-59904-931-1.ch059

Badr, K. B., & Ahmad, M. N. (2013). Managing lessons learned: A

comparative study of lessons learned systems . In Nazir Ahmad, M.,
Colomb, R., & Abdullah, M. (Eds.), Ontology-based applications for
enterprise systems and knowledge management (pp. 224–245).
Hershey, PA: IGI Global. doi:10.4018/978-1-4666-1993-7.ch013

Bakshi, K. (2014). Technologies for big data . In Hu, W., &

Kaabouch, N. (Eds.), Big data management, technologies, and
applications (pp. 1–22). Hershey, PA: IGI Global. doi:10.4018/978-1-
4666-4699-5.ch001

Ballou, D. P., Belardo, S., & Pazer, H. L. (2012). A project staffing

model to enhance the effectiveness of knowledge transfer in the
requirements planning phase for multi-project environments . In
Jennex, M. (Ed.), Conceptual models and outcomes of advancing
knowledge management: New technologies (pp. 77–98). Hershey, PA:
IGI Global. doi:10.4018/978-1-4666-0035-5.ch005

Baporikar, N. (2011). Knowledge management and entrepreneurship

cases in India . In Al-Shammari, M. (Ed.), Knowledge management in
emerging economies: Social, organizational and cultural
implementation (pp. 325–346). Hershey, PA: IGI Global.
doi:10.4018/978-1-61692-886-5.ch020

Baporikar, N. (2014). Knowledge management initiatives in Indian

public sector . In Al-Bastaki, Y., & Shajera, A. (Eds.), Building a
competitive public sector with knowledge management strategy (pp.
53–89). Hershey, PA: IGI Global. doi:10.4018/978-1-4666-4434-
2.ch002

Baporikar, N. (2014). Organizational barriers and facilitators in

embedding knowledge strategy . In Chilton, M., & Bloodgood, J.
(Eds.), Knowledge management and competitive advantage: Issues
and potential solutions (pp. 149–173). Hershey, PA: IGI Global.
doi:10.4018/978-1-4666-4679-7.ch009

Barioni, M. C., Kaster, D. D., Razente, H. L., Traina, A. J., & Júnior,
C. T. (2011). Querying multimedia data by similarity in relational
DBMS . In Yan, L., & Ma, Z. (Eds.), Advanced database query
systems: Techniques, applications and technologies (pp. 323–359).
Hershey, PA: IGI Global. doi:10.4018/978-1-60960-475-2.ch014

Baroni de Carvalho, R., & Tavares Ferreira, M. A. (2011). Knowledge

management software . In Schwartz, D., & Te’eni, D. (Eds.),
Encyclopedia of knowledge management (2nd ed.; pp. 738–749).
Hershey, PA: IGI Global. doi:10.4018/978-1-59904-931-1.ch072

Barroso, A. C., Ricciardi, R. I., & Junior, J. A. (2012). Web 2.0 and
project management: Reviewing the change path and discussing a few
cases . In Boughzala, I., & Dudezert, A. (Eds.), Knowledge
management 2.0: Organizational models and enterprise strategies (pp.
164–189). Hershey, PA: IGI Global. doi:10.4018/978-1-61350-195-
5.ch009

Baskaran, V., Naguib, R., Guergachi, A., Bali, R., & Arochen, H.
(2011). Does knowledge management really work? A case study in
the breast cancer screening domain . In Eardley, A., & Uden, L.
(Eds.), Innovative knowledge management: Concepts for
organizational creativity and collaborative design (pp. 177–189).
Hershey, PA: IGI Global. doi:10.4018/978-1-60566-701-0.ch010

Bebensee, T., Helms, R., & Spruit, M. (2012). Exploring the impact of
web 2.0 on knowledge management . In Boughzala, I., & Dudezert, A.
(Eds.), Knowledge management 2.0: Organizational models and
enterprise strategies (pp. 17–43). Hershey, PA: IGI Global.
doi:10.4018/978-1-61350-195-5.ch002

Becerra-Fernandez, I., & Sabherwal, R. (2011). The role of

information and communication technologies in knowledge
management: A classification of knowledge management systems . In
Schwartz, D., & Te’eni, D. (Eds.), Encyclopedia of knowledge
management (2nd ed.; pp. 1410–1418). Hershey, PA: IGI Global.
doi:10.4018/978-1-59904-931-1.ch134

Benbya, H. (2013). Valuing knowledge-based initiatives: What we

know and what we don’t know . In Jennex, M. (Ed.), Dynamic models
for knowledge-driven organizations (pp. 1–15). Hershey, PA: IGI
Global. doi:10.4018/978-1-4666-2485-6.ch001

Berends, H., van der Bij, H., & Weggeman, M. (2011). Knowledge
integration . In Schwartz, D., & Te’eni, D. (Eds.), Encyclopedia of
knowledge management (2nd ed.; pp. 581–590). Hershey, PA: IGI
Global. doi:10.4018/978-1-59904-931-1.ch056

Berger, H., & Beynon-Davies, P. (2011). Knowledge-based diffusion

in practice: A case study experience . In Eardley, A., & Uden, L.
(Eds.), Innovative knowledge management: Concepts for
organizational creativity and collaborative design (pp. 40–55).
Hershey, PA: IGI Global. doi:10.4018/978-1-60566-701-0.ch003
Berio, G., Di Leva, A., Harzallah, M., & Sacco, G. M. (2012).
Competence management over social networks through dynamic
taxonomies . In Boughzala, I., & Dudezert, A. (Eds.), Knowledge
management 2.0: Organizational models and enterprise strategies (pp.
103–120). Hershey, PA: IGI Global. doi:10.4018/978-1-61350-195-
5.ch006

Bhatt, S., Chaudhary, S., & Bhise, M. (2013). Migration of data

between cloud and non-cloud datastores . In Ionita, A., Litoiu, M., &
Lewis, G. (Eds.), Migrating legacy applications: Challenges in service
oriented architecture and cloud computing environments (pp. 206–
225). Hershey, PA: IGI Global. doi:10.4018/978-1-4666-2488-
7.ch009

Bloodgood, J. M., Chilton, M. A., & Bloodgood, T. C. (2014). The

effect of knowledge transfer motivation, receiver capability, and
motivation on organizational performance . In Chilton, M., &
Bloodgood, J. (Eds.), Knowledge management and competitive
advantage: Issues and potential solutions (pp. 232–242). Hershey, PA:
IGI Global. doi:10.4018/978-1-4666-4679-7.ch013

Boersma, K., & Kingma, S. (2011). Organizational learning

facilitation with intranet (2.0): A socio-cultural approach . In
Schwartz, D., & Te’eni, D. (Eds.), Encyclopedia of knowledge
management (2nd ed., pp. 1280–1289). Hershey, PA: IGI Global.
doi:10.4018/978-1-59904-931-1.ch122

Bond, P. L. (2010). Toward a living systems framework for unifying

technology and knowledge management, organizational, cultural and
economic change . In Harorimana, D. (Ed.), Cultural implications of
knowledge sharing, management and transfer: Identifying competitive
advantage (pp. 108–132). Hershey, PA: IGI Global. doi:10.4018/978-
1-60566-790-4.ch006
Bordogna, G., Bucci, F., Carrara, P., Pepe, M., & Rampini, A. (2011).
Flexible querying of imperfect temporal metadata in spatial data
infrastructures . In Yan, L., & Ma, Z. (Eds.), Advanced database
query systems: Techniques, applications and technologies (pp. 140–
159). Hershey, PA: IGI Global. doi:10.4018/978-1-60960-475-
2.ch006

Boughzala, I. (2012). Collaboration 2.0 through the new organization

(2.0) transformation . In Boughzala, I., & Dudezert, A. (Eds.),
Knowledge management 2.0: Organizational models and enterprise
strategies (pp. 1–16). Hershey, PA: IGI Global. doi:10.4018/978-1-
61350-195-5.ch001

Bratianu, C. (2011). A new perspective of the intellectual capital

dynamics in organizations . In Vallejo-Alonso, B., Rodriguez-
Castellanos, A., & Arregui-Ayastuy, G. (Eds.), Identifying,
measuring, and valuing knowledge-based intangible assets: New
perspectives (pp. 1–21). Hershey, PA: IGI Global. doi:10.4018/978-1-
60960-054-9.ch001

Bratianu, C. (2011). Universities as knowledge-intensive learning

organizations . In Eardley, A., & Uden, L. (Eds.), Innovative
knowledge management: Concepts for organizational creativity and
collaborative design (pp. 1–17). Hershey, PA: IGI Global.
doi:10.4018/978-1-60566-701-0.ch001

Breu, K., Ward, J., & Murray, P. (2000). Success factors in leveraging
the corporate information and knowledge resource through intranets .
In Malhotra, Y. (Ed.), Knowledge management and virtual
organizations (pp. 306–320). Hershey, PA: IGI Global.
doi:10.4018/978-1-930708-65-5.ch016

Briones-Peñalver, A., & Poças-Rascão, J. (2014). Information

technologies (ICT), network organizations, and information systems
for business cooperation: A focus on organization and strategic
knowledge management . In Jamil, G., Malheiro, A., & Ribeiro, F.
(Eds.), Rethinking the conceptual base for new practical applications
in information value and quality (pp. 324–348). Hershey, PA: IGI
Global. doi:10.4018/978-1-4666-4562-2.ch015

Brock, J. K., & Zhou, Y. J. (2011). MNE knowledge management

across borders and ICT . In Schwartz, D., & Te’eni, D. (Eds.),
Encyclopedia of knowledge management (2nd ed.; pp. 1136–1148).
Hershey, PA: IGI Global. doi:10.4018/978-1-59904-931-1.ch109

Brunet-Thornton, R., & Bureš, V. (2011). Meeting Czech knowledge

management challenges head-on: KM-Be.At-It . In Al-Shammari, M.
(Ed.), Knowledge management in emerging economies: Social,
organizational and cultural implementation (pp. 20–46). Hershey, PA:
IGI Global. doi:10.4018/978-1-61692-886-5.ch002

Bucher, T., & Dinter, B. (2012). Situational method engineering to

support process-oriented information logistics: Identification of
development situations. Journal of Database Management , 23(1), 31–
48. doi:10.4018/jdm.2012010102

Burstein, F., & Linger, H. (2011). Task-based knowledge

management approach . In Schwartz, D., & Te’eni, D. (Eds.),
Encyclopedia of knowledge management (2nd ed.; pp. 1479–1489).
Hershey, PA: IGI Global. doi:10.4018/978-1-59904-931-1.ch141

Butler, T. (2011). Anti-foundational knowledge management . In

Schwartz, D., & Te’eni, D. (Eds.), Encyclopedia of knowledge
management (2nd ed.; pp. 1–11). Hershey, PA: IGI Global.
doi:10.4018/978-1-59904-931-1.ch001

Butler, T., & Murphy, C. (2011). Work and knowledge . In Schwartz,

D., & Te’eni, D. (Eds.), Encyclopedia of knowledge management
(2nd ed.; pp. 1556–1566). Hershey, PA: IGI Global. doi:10.4018/978-
1-59904-931-1.ch148

Cabrilo, S., & Grubic-Nesic, L. (2013). The role of creativity,

innovation, and invention in knowledge management . In Buckley, S.,
& Jakovljevic, M. (Eds.), Knowledge management innovations for
interdisciplinary education: Organizational applications (pp. 207–
232). Hershey, PA: IGI Global. doi:10.4018/978-1-4666-1969-
2.ch011

Cabrita, M. D., Machado, V. C., & Grilo, A. (2010). Intellectual

capital: How knowledge creates value . In O’Brien, E., Clifford, S., &
Southern, M. (Eds.), Knowledge management for process,
organizational and marketing innovation: Tools and methods (pp.
237–252). Hershey, PA: IGI Global. doi:10.4018/978-1-61520-829-
6.ch015

Cagliero, L., & Fiori, A. (2013). Knowledge discovery from online

communities . In Data mining: Concepts, methodologies, tools, and
applications (pp. 1230–1252). Hershey, PA: IGI Global.
doi:10.4018/978-1-4666-2455-9.ch063

Camisón-Zornoza, C., & Boronat-Navarro, M. (2010). Linking

exploration and exploitation capabilities with the process of
knowledge development and with organizational facilitators . In Russ,
M. (Ed.), Knowledge management strategies for business
development (pp. 159–179). Hershey, PA: IGI Global.
doi:10.4018/978-1-60566-348-7.ch008

Carneiro, A. (2010). Change knowledge management: Transforming a

ghost community into a real asset . In O’Brien, E., Clifford, S., &
Southern, M. (Eds.), Knowledge management for process,
organizational and marketing innovation: Tools and methods (pp.
120–132). Hershey, PA: IGI Global. doi:10.4018/978-1-61520-829-
6.ch007

Carrillo, F. J. (2010). Knowledge-based value generation . In

Metaxiotis, K., Carrillo, F., & Yigitcanlar, T. (Eds.), Knowledge-
based development for cities and societies: Integrated multi-level
approaches (pp. 1–16). Hershey, PA: IGI Global. doi:10.4018/978-1-
61520-721-3.ch001

Cartelli, A. (2012). Frameworks for the benchmarking of digital and

knowledge management best practice in SME and organizations . In
Cartelli, A. (Ed.), Current trends and future practices for digital
literacy and competence (pp. 166–175). Hershey, PA: IGI Global.
doi:10.4018/978-1-4666-0903-7.ch015

Castellano, G., Fanelli, A. M., & Torsello, M. A. (2010). Soft

computing techniques in content-based multimedia information
retrieval . In Anbumani, K., & Nedunchezhian, R. (Eds.), Soft
computing applications for database technologies: Techniques and
issues (pp. 170–192). Hershey, PA: IGI Global. doi:10.4018/978-1-
60566-814-7.ch010

Chalkiti, K., & Carson, D. (2010). Knowledge cultures, competitive

advantage and staff turnover in hospitality in Australia’s northern
territory . In Harorimana, D. (Ed.), Cultural implications of
knowledge sharing, management and transfer: Identifying competitive
advantage (pp. 203–229). Hershey, PA: IGI Global. doi:10.4018/978-
1-60566-790-4.ch010

Chang, W., & Li, S. (2011). Deploying knowledge management in

R&D workspaces . In Eardley, A., & Uden, L. (Eds.), Innovative
knowledge management: Concepts for organizational creativity and
collaborative design (pp. 56–76). Hershey, PA: IGI Global.
doi:10.4018/978-1-60566-701-0.ch004
Chawla, D., & Joshi, H. (2013). Impact of knowledge management
dimensions on learning organization: Comparison across business
excellence awarded and non-awarded indian organizations . In Jennex,
M. (Ed.), Dynamic models for knowledge-driven organizations (pp.
145–162). Hershey, PA: IGI Global. doi:10.4018/978-1-4666-2485-
6.ch008

Chen, E. (2012). Web 2.0 social networking technologies and

strategies for knowledge management . In Boughzala, I., & Dudezert,
A. (Eds.), Knowledge management 2.0: Organizational models and
enterprise strategies (pp. 84–102). Hershey, PA: IGI Global.
doi:10.4018/978-1-61350-195-5.ch005

Chihara, K., & Nakamori, Y. (2013). Clarification of abilities and

qualities of knowledge coordinators: The case of regional
revitalization projects . In Yang, G. (Ed.), Multidisciplinary studies in
knowledge and systems science (pp. 1–17). Hershey, PA: IGI Global.
doi:10.4018/978-1-4666-3998-0.ch001

Christidis, K., Papailiou, N., Apostolou, D., & Mentzas, G. (2011).

Semantic interfaces for personal and social knowledge work.
International Journal of Knowledge-Based Organizations , 1(1), 61–
77. doi:10.4018/ijkbo.2011010104

Christidis, K., Papailiou, N., Apostolou, D., & Mentzas, G. (2013).

Semantic interfaces for personal and social knowledge work . In
Wang, J. (Ed.), Intelligence methods and systems advancements for
knowledge-based business (pp. 213–230). Hershey, PA: IGI Global.
doi:10.4018/978-1-4666-1873-2.ch012

Chua, C. E., Storey, V. C., & Chiang, R. H. (2012). Knowledge

representation: A conceptual modeling approach. Journal of Database
Management , 23(1), 1–30. doi:10.4018/jdm.2012010101
Clinton, M. S., Merritt, K. L., & Murray, S. R. (2011). Facilitating
knowledge transfer and the achievement of competitive advantage
with corporate universities: An exploratory model based on media
richness and type of knowledge to be transferred . In Jennex, M. (Ed.),
Global aspects and cultural perspectives on knowledge management:
Emerging dimensions (pp. 329–345). Hershey, PA: IGI Global.
doi:10.4018/978-1-60960-555-1.ch020

Colomb, R. M. (2013). Representation of action is a primary

requirement in ontologies for interoperating information systems . In
Nazir Ahmad, M., Colomb, R., & Abdullah, M. (Eds.), Ontology-
based applications for enterprise systems and knowledge management
(pp. 68–76). Hershey, PA: IGI Global. doi:10.4018/978-1-4666-1993-
7.ch004

Colucci, S., Di Noia, T., Di Sciascio, E., Donini, F. M., & Mongiello,
M. (2011). Description logic-based resource retrieval . In Schwartz,
D., & Te’eni, D. (Eds.), Encyclopedia of knowledge management
(2nd ed.; pp. 185–197). Hershey, PA: IGI Global. doi:10.4018/978-1-
59904-931-1.ch018

Connell, N. A. (2011). Organisational storytelling . In Schwartz, D., &

Te’eni, D. (Eds.), Encyclopedia of knowledge management (2nd ed.;
pp. 1261–1269). Hershey, PA: IGI Global. doi:10.4018/978-1-59904-
931-1.ch120

Cooper, L. P., & Rober, M. B. (2012). Moving wikis behind the

firewall: Intrapedias and work-wikis . In Boughzala, I., & Dudezert,
A. (Eds.), Knowledge Management 2.0: Organizational Models and
Enterprise Strategies (pp. 44–63). Hershey, PA: IGI Global.
doi:10.4018/978-1-61350-195-5.ch003

Corallo, A., De Maggio, M., & Margherita, A. (2011). Knowledge

democracy as the new mantra in product innovation: A framework of
processes and competencies . In Eardley, A., & Uden, L. (Eds.),
Innovative knowledge management: Concepts for organizational
creativity and collaborative design (pp. 141–156). Hershey, PA: IGI
Global. doi:10.4018/978-1-60566-701-0.ch008

Costa, G. (2011). Knowledge worker fair compensation: Ethical

issues and social dilemmas . In Morais da Costa, G. (Ed.), Ethical
issues and social dilemmas in knowledge management:
Organizational innovation (pp. 215–231). Hershey, PA: IGI Global.
doi:10.4018/978-1-61520-873-9.ch013

Costello, R. (2014). Evaluating e-learning from an end user

perspective . In Pańkowska, M. (Ed.), Frameworks of IT prosumption
for business development (pp. 259–283). Hershey, PA: IGI Global.
doi:10.4018/978-1-4666-4313-0.ch017

Crasso, M., Zunino, A., & Campo, M. (2013). A survey of approaches

to web service discovery in service-oriented architectures . In Siau, K.
(Ed.), Innovations in database design, web applications, and
information systems management (pp. 107–138). Hershey, PA: IGI
Global. doi:10.4018/978-1-4666-2044-5.ch005

Croasdell, D., & Wang, Y. K. (2011). Virtue-nets . In Schwartz, D., &

Te’eni, D. (Eds.), Encyclopedia of knowledge management (2nd ed.;
pp. 1545–1555). Hershey, PA: IGI Global. doi:10.4018/978-1-59904-
931-1.ch147

Cucchiara, S., Ligorio, M. B., & Fujita, N. (2014). Understanding

online discourse strategies for knowledge building through social
network analysis . In Lim, H., & Sudweeks, F. (Eds.), Innovative
methods and technologies for electronic discourse analysis (pp. 42–
62). Hershey, PA: IGI Global. doi:10.4018/978-1-4666-4426-7.ch003

Cudanov, M., & Kirchner, K. (2011). Knowledge management in

high-growth companies: A case study in Serbia . In Al-Shammari, M.
(Ed.), Knowledge management in emerging economies: Social,
organizational and cultural implementation (pp. 227–248). Hershey,
PA: IGI Global. doi:10.4018/978-1-61692-886-5.ch014

Cuel, R., Bouquet, P., & Bonifacio, M. (2011). Distributed knowledge

management . In Schwartz, D., & Te’eni, D. (Eds.), Encyclopedia of
knowledge management (2nd ed.; pp. 198–208). Hershey, PA: IGI
Global. doi:10.4018/978-1-59904-931-1.ch019

Daidj, N. (2012). The evolution of KM practices: The case of the

Renault-Nissan international strategic alliance . In Boughzala, I., &
Dudezert, A. (Eds.), Knowledge management 2.0: Organizational
models and enterprise strategies (pp. 190–213). Hershey, PA: IGI
Global. doi:10.4018/978-1-61350-195-5.ch010

Daniel, B. K., Zapata-Rivera, J., & McCalla, G. I. (2010). A Bayesian

belief network methodology for modeling social systems in virtual
communities: Opportunities for database technologies . In Anbumani,
K., & Nedunchezhian, R. (Eds.), Soft computing applications for
database technologies: Techniques and issues (pp. 125–152). Hershey,
PA: IGI Global. doi:10.4018/978-1-60566-814-7.ch008

Darchen, S., & Tremblay, D. (2010). Attracting and retaining

knowledge workers: The impact of quality of place in the case of
Montreal . In Metaxiotis, K., Carrillo, F., & Yigitcanlar, T. (Eds.),
Knowledge-based development for cities and societies: Integrated
multi-level approaches (pp. 42–58). Hershey, PA: IGI Global.
doi:10.4018/978-1-61520-721-3.ch003

Davenport, D. L., & Hosapple, C. W. (2011). Knowledge

organizations . In Schwartz, D., & Te’eni, D. (Eds.), Encyclopedia of
knowledge management (2nd ed.; pp. 822–832). Hershey, PA: IGI
Global. doi:10.4018/978-1-59904-931-1.ch079
Davenport, D. L., & Hosapple, C. W. (2011). Social capital
knowledge . In Schwartz, D., & Te’eni, D. (Eds.), Encyclopedia of
knowledge management (2nd ed.; pp. 1448–1459). Hershey, PA: IGI
Global. doi:10.4018/978-1-59904-931-1.ch138

De Maggio, M., Del Vecchio, P., Elia, G., & Grippa, F. (2011). An
ICT-based network of competence centres for developing intellectual
capital in the Mediterranean area . In Al Ajeeli, A., & Al-Bastaki, Y.
(Eds.), Handbook of research on e-services in the public sector: E-
government strategies and advancements (pp. 164–181). Hershey, PA:
IGI Global. doi:10.4018/978-1-61520-789-3.ch014

Delbaere, M., Di Zhang, D., Bruning, E. R., & Sivaramakrishnan, S.

(2014). Knowledge management and the roles it plays in achieving
superior performance . In Chilton, M., & Bloodgood, J. (Eds.),
Knowledge management and competitive advantage: Issues and
potential solutions (pp. 90–108). Hershey, PA: IGI Global.
doi:10.4018/978-1-4666-4679-7.ch006

Delgado-Verde, M., & Cruz-González, J. (2010). An intellectual

capital-based view of technological innovation . In López Sáez, P.,
Castro, G., Navas López, J., & Delgado Verde, M. (Eds.), Intellectual
capital and technological innovation: Knowledge-based theory and
practice (pp. 166–193). Hershey, PA: IGI Global. doi:10.4018/978-1-
61520-875-3.ch008

Deltour, F., Plé, L., & Roussel, C. S. (2012). Knowledge sharing in

the age of web 2.0: A social capital perspective . In Boughzala, I., &
Dudezert, A. (Eds.), Knowledge management 2.0: Organizational
models and enterprise strategies (pp. 122–141). Hershey, PA: IGI
Global. doi:10.4018/978-1-61350-195-5.ch007

Derballa, V., & Pousttchi, K. (2011). Mobile technology for

knowledge management . In Schwartz, D., & Te’eni, D. (Eds.),
Encyclopedia of knowledge management (2nd ed.; pp. 1158–1166).
Hershey, PA: IGI Global. doi:10.4018/978-1-59904-931-1.ch111

Dieng-Kuntz, R. (2011). Corporate semantic webs . In Schwartz, D.,

& Te’eni, D. (Eds.), Encyclopedia of knowledge management (2nd
ed.; pp. 131–149). Hershey, PA: IGI Global. doi:10.4018/978-1-
59904-931-1.ch014

Diosteanu, A., Stellato, A., & Turbati, A. (2012). SODA: A service

oriented data acquisition framework . In Pazienza, M., & Stellato, A.
(Eds.), Semi-automatic ontology development: Processes and
resources (pp. 48–77). Hershey, PA: IGI Global. doi:10.4018/978-1-
4666-0188-8.ch003

Donate-Manzanares, M. J., Guadamillas-Gómez, F., & Sánchez de

Pablo, J. D. (2010). Strategic alliances and knowledge management
strategies: A case study . In Russ, M. (Ed.), Knowledge management
strategies for business development (pp. 240–260). Hershey, PA: IGI
Global. doi:10.4018/978-1-60566-348-7.ch011

Donnet, T., Keast, R., & Pickernell, D. (2010). Up the junction?

Exploiting knowledge-based development through supply chain and
SME cluster interactions . In Metaxiotis, K., Carrillo, F., &
Yigitcanlar, T. (Eds.), Knowledge-based development for cities and
societies: Integrated multi-level approaches (pp. 179–195). Hershey,
PA: IGI Global. doi:10.4018/978-1-61520-721-3.ch011

Douglas, I. (2011). Organizational needs analysis and knowledge

management . In Schwartz, D., & Te’eni, D. (Eds.), Encyclopedia of
knowledge management (2nd ed.; pp. 1290–1297). Hershey, PA: IGI
Global. doi:10.4018/978-1-59904-931-1.ch123

Edvardsson, I. R., & Oskarsson, G. K. (2013). Outsourcing in

knowledge-based service firms . In Wang, J. (Ed.), Intelligence
methods and systems advancements for knowledge-based business
(pp. 97–113). Hershey, PA: IGI Global. doi:10.4018/978-1-4666-
1873-2.ch006

Elenurm, T. (2013). Knowledge management and innovative learning

. In Buckley, S., & Jakovljevic, M. (Eds.), Knowledge management
innovations for interdisciplinary education: Organizational
applications (pp. 108–131). Hershey, PA: IGI Global.
doi:10.4018/978-1-4666-1969-2.ch006

Eppler, M. J., & Burkhard, R. A. (2011). Knowledge visualization . In

Schwartz, D., & Te’eni, D. (Eds.), Encyclopedia of knowledge
management (2nd ed.; pp. 987–999). Hershey, PA: IGI Global.
doi:10.4018/978-1-59904-931-1.ch094

Ergazakis, K., Metaxiotis, K., & Ergazakis, E. (2011). Exploring paths

towards knowledge cities developments: A research agenda . In
Schwartz, D., & Te’eni, D. (Eds.), Encyclopedia of knowledge
management (2nd ed.; pp. 288–297). Hershey, PA: IGI Global.
doi:10.4018/978-1-59904-931-1.ch028

Eri, Z. D., Abdullah, R., Jabar, M. A., Murad, M. A., & Talib, A. M.
(2013). Ontology-based virtual communities model for the knowledge
management system environment: Ontology design . In Nazir Ahmad,
M., Colomb, R., & Abdullah, M. (Eds.), Ontology-based applications
for enterprise systems and knowledge management (pp. 343–360).
Hershey, PA: IGI Global. doi:10.4018/978-1-4666-1993-7.ch019

Erickson, G. S. (2014). Government as a partner in knowledge

management: Lessons from the US freedom of information act . In Al-
Bastaki, Y., & Shajera, A. (Eds.), Building a competitive public sector
with knowledge management strategy (pp. 90–103). Hershey, PA: IGI
Global. doi:10.4018/978-1-4666-4434-2.ch003
Erickson, G. S., & Rothberg, H. N. (2011). Assessing knowledge
management needs: A strategic approach to developing knowledge.
International Journal of Knowledge Management , 7(3), 1–10.
doi:10.4018/jkm.2011070101

Erickson, G. S., & Rothberg, H. N. (2011). Protecting knowledge

assets . In Schwartz, D., & Te’eni, D. (Eds.), Encyclopedia of
knowledge management (2nd ed.; pp. 1336–1342). Hershey, PA: IGI
Global. doi:10.4018/978-1-59904-931-1.ch128

Erickson, G. S., & Rothberg, H. N. (2013). Assessing knowledge

management needs: A strategic approach to developing knowledge .
In Jennex, M. (Ed.), Dynamic models for knowledge-driven
organizations (pp. 180–189). Hershey, PA: IGI Global.
doi:10.4018/978-1-4666-2485-6.ch010

Evermann, J., & Wand, Y. (2011). Ontology based object-oriented

domain modeling: Representing behavior . In Siau, K. (Ed.),
Theoretical and practical advances in information systems
development: Emerging trends and approaches (pp. 37–60). Hershey,
PA: IGI Global. doi:10.4018/978-1-60960-521-6.ch003

Fadel, K. J., Durcikova, A., & Cha, H. S. (2011). An experiment of

information elaboration in mediated knowledge transfer . In Jennex,
M. (Ed.), Global aspects and cultural perspectives on knowledge
management: Emerging dimensions (pp. 311–328). Hershey, PA: IGI
Global. doi:10.4018/978-1-60960-555-1.ch019

Fazel-Zarandi, M., Fox, M. S., & Yu, E. (2013). Ontologies in

expertise finding systems: Modeling, analysis, and design . In Nazir
Ahmad, M., Colomb, R., & Abdullah, M. (Eds.), Ontology-based
applications for enterprise systems and knowledge management (pp.
158–177). Hershey, PA: IGI Global. doi:10.4018/978-1-4666-1993-
7.ch009
Ferri, F., & Grifoni, P. (2011). Sketching in knowledge creation and
management . In Schwartz, D., & Te’eni, D. (Eds.), Encyclopedia of
knowledge management (2nd ed.; pp. 1438–1447). Hershey, PA: IGI
Global. doi:10.4018/978-1-59904-931-1.ch137

Filho, C. G., Baroni de Carvalho, R., & Jamil, G. L. (2011). Market

knowledge management, innovation and product performance: Survey
in medium and large Brazilian industrial firms . In Jennex, M., &
Smolnik, S. (Eds.), Strategies for knowledge management success:
Exploring organizational efficacy (pp. 32–50). Hershey, PA: IGI
Global. doi:10.4018/978-1-60566-709-6.ch003

Fink, D., & Disterer, G. (2011). Knowledge management in

professional service firms . In Schwartz, D., & Te’eni, D. (Eds.),
Encyclopedia of knowledge management (2nd ed.; pp. 650–659).
Hershey, PA: IGI Global. doi:10.4018/978-1-59904-931-1.ch063

Fink, K. (2011). Process model for knowledge potential measurement

in SMEs . In Jennex, M., & Smolnik, S. (Eds.), Strategies for
knowledge management success: Exploring organizational efficacy
(pp. 91–105). Hershey, PA: IGI Global. doi:10.4018/978-1-60566-
709-6.ch006

Fink, K., & Ploder, C. (2011). Knowledge management toolkit for

SMEs . In Jennex, M. (Ed.), Global aspects and cultural perspectives
on knowledge management: Emerging dimensions (pp. 49–63).
Hershey, PA: IGI Global. doi:10.4018/978-1-60960-555-1.ch004

Flynn, R., & Marshall, V. (2014). The four levers for change in
knowledge management implementation . In Al-Bastaki, Y., &
Shajera, A. (Eds.), Building a competitive public sector with
knowledge management strategy (pp. 227–245). Hershey, PA: IGI
Global. doi:10.4018/978-1-4666-4434-2.ch010
Fortier, J., & Kassel, G. (2011). Organizational semantic webs . In
Schwartz, D., & Te’eni, D. (Eds.), Encyclopedia of knowledge
management (2nd ed.; pp. 1298–1307). Hershey, PA: IGI Global.
doi:10.4018/978-1-59904-931-1.ch124

Framinan, J. M., & Molina, J. M. (2010). An overview of enterprise

resource planning for intelligent enterprises . In Business information
systems: Concepts, methodologies, tools and applications (pp. 60–68).
Hershey, PA: IGI Global. doi:10.4018/978-1-61520-969-9.ch005

Franco, M., Di Virgilio, F., & Di Pietro, L. (2014). Management of

group knowledge and the role of E-WOM for business organizations .
In Chilton, M., & Bloodgood, J. (Eds.), Knowledge management and
competitive advantage: Issues and potential solutions (pp. 71–89).
Hershey, PA: IGI Global. doi:10.4018/978-1-4666-4679-7.ch005

Franke, U. J. (2000). The knowledge-based view (KBV) of the virtual

web, the virtual corporation and the net-broker . In Malhotra, Y. (Ed.),
Knowledge management and virtual organizations (pp. 20–42).
Hershey, PA: IGI Global. doi:10.4018/978-1-930708-65-5.ch002

Freivalds, D., & Lush, B. (2012). Thinking inside the grid: Selecting a
discovery system through the RFP process . In Popp, M., & Dallis, D.
(Eds.), Planning and implementing resource discovery tools in
academic libraries (pp. 104–121). Hershey, PA: IGI Global.
doi:10.4018/978-1-4666-1821-3.ch007

Frieß, M. R., Groh, G., Reinhardt, M., Forster, F., & Schlichter, J.
(2012). Context-aware creativity support for corporate open
innovation. International Journal of Knowledge-Based Organizations ,
2(1), 38–55. doi:10.4018/ijkbo.2012010103

Fuller, C. M., & Wilson, R. L. (2011). Extracting knowledge from

neural networks . In Schwartz, D., & Te’eni, D. (Eds.), Encyclopedia
of knowledge management (2nd ed.; pp. 320–330). Hershey, PA: IGI
Global. doi:10.4018/978-1-59904-931-1.ch031

Furquim, T. D., & do Amaral, S. A. (2011). Knowledge management

practices in brazilian software organizations: The case of SERPRO .
In Al-Shammari, M. (Ed.), Knowledge management in emerging
economies: Social, organizational and cultural implementation (pp.
213–226). Hershey, PA: IGI Global. doi:10.4018/978-1-61692-886-
5.ch013

Gaál, Z., Szabó, L., Obermayer-Kovács, N., Kovács, Z., & Csepregi,
A. (2011). Knowledge management profile: An innovative approach
to map knowledge management practice . In Eardley, A., & Uden, L.
(Eds.), Innovative knowledge management: Concepts for
organizational creativity and collaborative design (pp. 253–263).
Hershey, PA: IGI Global. doi:10.4018/978-1-60566-701-0.ch016

Ganguly, A., Mostashari, A., & Mansouri, M. (2013). Measuring

knowledge management/knowledge sharing (KM/KS) efficiency and
effectiveness in enterprise networks . In Jennex, M. (Ed.), Dynamic
models for knowledge-driven organizations (pp. 318–336). Hershey,
PA: IGI Global. doi:10.4018/978-1-4666-2485-6.ch019

Gaumand, C., Chapdaniel, A., & Dudezert, A. (2012). Strategic

knowledge management system framework for supply chain at an
intra-organizational level . In Boughzala, I., & Dudezert, A. (Eds.),
Knowledge management 2.0: Organizational models and enterprise
strategies (pp. 142–163). Hershey, PA: IGI Global. doi:10.4018/978-
1-61350-195-5.ch008

Ghazali, R., & Zakaria, N. H. (2013). Knowledge management

processes in enterprise systems: A systematic literature review . In
Nazir Ahmad, M., Colomb, R., & Abdullah, M. (Eds.), Ontology-
based applications for enterprise systems and knowledge management
(pp. 1–24). Hershey, PA: IGI Global. doi:10.4018/978-1-4666-1993-
7.ch001

Ghosh, B. (2011). Cross-cultural knowledge management practices to

support offshore outsourcing . In Al-Shammari, M. (Ed.), Knowledge
management in emerging economies: Social, organizational and
cultural implementation (pp. 249–260). Hershey, PA: IGI Global.
doi:10.4018/978-1-61692-886-5.ch015

Gohil, U., Carrillo, P., Ruikar, K., & Anumba, C. (2013).

Development of a business process model for a project-based service
organisation. International Journal of Knowledge-Based
Organizations , 3(1), 37–56. doi:10.4018/ijkbo.2013010103

Goldsmith, R. E., & Pillai, K. G. (2011). Knowledge calibration and

knowledge management . In Schwartz, D., & Te’eni, D. (Eds.),
Encyclopedia of knowledge management (2nd ed.; pp. 497–505).
Hershey, PA: IGI Global. doi:10.4018/978-1-59904-931-1.ch048

Gomes de Andrade, F., & Baptista, C. D. (2013). An ontology-based

approach to support information discovery in spatial data
infrastructures . In Rückemann, C. (Ed.), Integrated information and
computing systems for natural, spatial, and social sciences (pp. 369–
387). Hershey, PA: IGI Global. doi:10.4018/978-1-4666-2190-
9.ch018

Gonçalo, C. R., & Jacques, E. J. (2010). Best practices of knowledge

strategy in hospitals: A contextual perspective based on the
implementation of medical protocols . In Harorimana, D. (Ed.),
Cultural implications of knowledge sharing, management and transfer:
Identifying competitive advantage (pp. 180–202). Hershey, PA: IGI
Global. doi:10.4018/978-1-60566-790-4.ch009

Górniak-Kocikowska, K. (2011). Knowledge management and

democracy: A critical review of some moral issues and social
dilemmas . In Morais da Costa, G. (Ed.), Ethical issues and social
dilemmas in knowledge management: Organizational innovation (pp.
28–44). Hershey, PA: IGI Global. doi:10.4018/978-1-61520-873-
9.ch003

Gottschalk, P. (2014). Police knowledge management strategy . In

Chilton, M., & Bloodgood, J. (Eds.), Knowledge management and
competitive advantage: Issues and potential solutions (pp. 202–220).
Hershey, PA: IGI Global. doi:10.4018/978-1-4666-4679-7.ch011

Goudos, S. K., Peristeras, V., & Tarabanis, K. (2010). Application of

semantic web technology in e-business: Case studies in public domain
data knowledge representation . In Business information systems:
Concepts, methodologies, tools and applications (pp. 1223–1233).
Hershey, PA: IGI Global. doi:10.4018/978-1-61520-969-9.ch075

Govindarajan, M., & Chandrasekaran, R. (2012). A hybrid multilayer

perceptron neural network for direct marketing. International Journal
of Knowledge-Based Organizations , 2(3), 63–73.
doi:10.4018/ijkbo.2012070104

Grant, J., & Minker, J. (2011). Logic and knowledge bases . In

Schwartz, D., & Te’eni, D. (Eds.), Encyclopedia of knowledge
management (2nd ed.; pp. 1022–1033). Hershey, PA: IGI Global.
doi:10.4018/978-1-59904-931-1.ch097

Green, A. (2011). Engineering business reasoning, analytics and

intelligence network (E-BRAIN): A new approach to intangible asset
valuation based on Einstein’s perspective . In Vallejo-Alonso, B.,
Rodriguez-Castellanos, A., & Arregui-Ayastuy, G. (Eds.), Identifying,
measuring, and valuing knowledge-based intangible assets: New
perspectives (pp. 232–253). Hershey, PA: IGI Global.
doi:10.4018/978-1-60960-054-9.ch011
Greenaway, K. E., & Vuong, D. C. (2010). Taking charities seriously:
A call for focused knowledge management research. International
Journal of Knowledge Management , 6(4), 87–97.
doi:10.4018/jkm.2010100105

Greenaway, K. E., & Vuong, D. C. (2012). Taking charities seriously:

A call for focused knowledge management research . In Jennex, M.
(Ed.), Conceptual models and outcomes of advancing knowledge
management: New technologies (pp. 333–344). Hershey, PA: IGI
Global. doi:10.4018/978-1-4666-0035-5.ch017

Gunjal, B., Gaitanou, P., & Yasin, S. (2012). Social networks and
knowledge management: An explorative study in library systems . In
Boughzala, I., & Dudezert, A. (Eds.), Knowledge management 2.0:
Organizational models and enterprise strategies (pp. 64–83). Hershey,
PA: IGI Global. doi:10.4018/978-1-61350-195-5.ch004

Habhab-Rave, S. (2010). Knowledge management in SMEs: A

mixture of innovation, marketing and ICT: Analysis of two case
studies . In O’Brien, E., Clifford, S., & Southern, M. (Eds.),
Knowledge management for process, organizational and marketing
innovation: Tools and methods (pp. 183–194). Hershey, PA: IGI
Global. doi:10.4018/978-1-61520-829-6.ch011

Habicht, H., Möslein, K. M., & Reichwald, R. (2012). Open

innovation maturity. International Journal of Knowledge-Based
Organizations , 2(1), 92–111. doi:10.4018/ijkbo.2012010106

Hamburg, I., & Hall, T. (2010). Readiness for knowledge

management, methods and environments for innovation . In O’Brien,
E., Clifford, S., & Southern, M. (Eds.), Knowledge management for
process, organizational and marketing innovation: Tools and methods
(pp. 1–15). Hershey, PA: IGI Global. doi:10.4018/978-1-61520-829-
6.ch001
Hamza, S. E. (2011). Capturing tacit knowledge from transient
workers: Improving the organizational competitiveness . In Jennex,
M. (Ed.), Global aspects and cultural perspectives on knowledge
management: Emerging dimensions (pp. 172–188). Hershey, PA: IGI
Global. doi:10.4018/978-1-60960-555-1.ch012

Harorimana, D. (2010). Knowledge, culture, and cultural impact on

knowledge management: Some lessons for researchers and
practitioners . In Harorimana, D. (Ed.), Cultural implications of
knowledge sharing, management and transfer: Identifying competitive
advantage (pp. 48–59). Hershey, PA: IGI Global. doi:10.4018/978-1-
60566-790-4.ch003

Hasan, H. (2011). Formal and emergent standards in KM . In

Schwartz, D., & Te’eni, D. (Eds.), Encyclopedia of knowledge
management (2nd ed.; pp. 331–342). Hershey, PA: IGI Global.
doi:10.4018/978-1-59904-931-1.ch032

He, G., Xue, G., Yu, K., & Yao, S. (2013). Business process
modeling: Analysis and evaluation . In Lu, Z. (Ed.), Design,
performance, and analysis of innovative information retrieval (pp.
382–393). Hershey, PA: IGI Global. doi:10.4018/978-1-4666-1975-
3.ch027

Heiman, B. A., & Hurmelinna-Laukkanen, P. (2010). Problem finding

and solving: A knowledge-based view of managing innovation . In
López Sáez, P., Castro, G., Navas López, J., & Delgado Verde, M.
(Eds.), Intellectual capital and technological innovation: Knowledge-
based theory and practice (pp. 105–130). Hershey, PA: IGI Global.
doi:10.4018/978-1-61520-875-3.ch005

Hendriks, P. H. (2011). Organizational structure . In Schwartz, D., &

Te’eni, D. (Eds.), Encyclopedia of knowledge management (2nd ed.;
pp. 1308–1318). Hershey, PA: IGI Global. doi:10.4018/978-1-59904-
931-1.ch125

Hercheui, M. D. (2012). KMS for fostering behavior change: A case

study on Microsoft Hohm . In Boughzala, I., & Dudezert, A. (Eds.),
Knowledge management 2.0: Organizational models and enterprise
strategies (pp. 214–232). Hershey, PA: IGI Global. doi:10.4018/978-
1-61350-195-5.ch011

Hipkin, I. (2011). Perceptions of factors influencing knowledge-based

technology management in conflict areas . In Al-Shammari, M. (Ed.),
Knowledge management in emerging economies: Social,
organizational and cultural implementation (pp. 294–307). Hershey,
PA: IGI Global. doi:10.4018/978-1-61692-886-5.ch018

Hofer, F. (2011). Knowledge transfer between academia and industry

. In Schwartz, D., & Te’eni, D. (Eds.), Encyclopedia of knowledge
management (2nd ed.; pp. 977–986). Hershey, PA: IGI Global.
doi:10.4018/978-1-59904-931-1.ch093

Holjevac, I. A., Crnjar, K., & Hrgovic, A. V. (2013). Knowledge

management and quality in Croatian tourism . In Buckley, S., &
Jakovljevic, M. (Eds.), Knowledge management innovations for
interdisciplinary education: Organizational applications (pp. 178–
192). Hershey, PA: IGI Global. doi:10.4018/978-1-4666-1969-
2.ch009

Holsapple, C. W., & Joshi, K. D. (2011). Knowledge management

ontology . In Schwartz, D., & Te’eni, D. (Eds.), Encyclopedia of
knowledge management (2nd ed.; pp. 704–711). Hershey, PA: IGI
Global. doi:10.4018/978-1-59904-931-1.ch068

Holsapple, C. W., & Oh, J. (2014). Reactive and proactive dynamic

capabilities: Using the knowledge chain theory of competitiveness . In
Chilton, M., & Bloodgood, J. (Eds.), Knowledge management and
competitive advantage: Issues and potential solutions (pp. 1–19).
Hershey, PA: IGI Global. doi:10.4018/978-1-4666-4679-7.ch001

Huang, A., Xiao, J., & Wang, S. (2013). A combined forecast method
integrating contextual knowledge . In Yang, G. (Ed.),
Multidisciplinary studies in knowledge and systems science (pp. 274–
290). Hershey, PA: IGI Global. doi:10.4018/978-1-4666-3998-
0.ch019

Huff, C. (2011). What does knowledge have to do with ethics? In

Morais da Costa, G. (Ed.), Ethical issues and social dilemmas in
knowledge management: Organizational innovation (pp. 17–27).
Hershey, PA: IGI Global. doi:10.4018/978-1-61520-873-9.ch002

Hürster, W., Wilbois, T., & Chaves, F. (2010). An integrated systems

approach for early warning and risk management systems.
International Journal of Information Technologies and Systems
Approach , 3(2), 46–56. doi:10.4018/jitsa.2010070104

Iyer, S. R., Sharda, R., Biros, D., Lucca, J., & Shimp, U. (2011).
Organization of lessons learned knowledge: A taxonomy and
implementation . In Jennex, M. (Ed.), Global aspects and cultural
perspectives on knowledge management: Emerging dimensions (pp.
190–209). Hershey, PA: IGI Global. doi:10.4018/978-1-60960-555-
1.ch013

Jacobson, C. M. (2011). Knowledge sharing between individuals . In

Schwartz, D., & Te’eni, D. (Eds.), Encyclopedia of knowledge
management (2nd ed.; pp. 924–934). Hershey, PA: IGI Global.
doi:10.4018/978-1-59904-931-1.ch088

Jakovljevic, M. (2013). A conceptual model of creativity, invention,

and innovation (MCII) for entrepreneurial engineers . In Buckley, S.,
& Jakovljevic, M. (Eds.), Knowledge management innovations for
interdisciplinary education: Organizational applications (pp. 66–87).
Hershey, PA: IGI Global. doi:10.4018/978-1-4666-1969-2.ch004

Jasimuddin, S. M., Connell, N., & Klein, J. H. (2011). Understanding

organizational memory . In Schwartz, D., & Te’eni, D. (Eds.),
Encyclopedia of knowledge management (2nd ed.; pp. 1536–1544).
Hershey, PA: IGI Global. doi:10.4018/978-1-59904-931-1.ch146

Jennex, M. E. (2010). Do organizational memory and information

technology interact to affect organizational information needs and
provision? In Ubiquitous developments in knowledge management:
Integrations and trends (pp. 1–20). Hershey, PA: IGI Global.
doi:10.4018/978-1-60566-954-0

Jennex, M. E. (2010). Knowledge sharing model of 24-hour

knowledge factory . In Ubiquitous developments in knowledge
management: Integrations and trends (pp. 141–154). Hershey, PA: IGI
Global. doi:10.4018/978-1-60566-954-0

Jennex, M. E. (2010). Operationalizing knowledge sharing for

informers . In Ubiquitous developments in knowledge management:
Integrations and trends (pp. 319–340). Hershey, PA: IGI Global.
doi:10.4018/978-1-60566-954-0

Jennex, M. E. (2010). Qualitative pre-processing for semantic search

of unstructured knowledge . In Ubiquitous developments in
knowledge management: Integrations and trends (pp. 252–263).
Hershey, PA: IGI Global. doi:10.4018/978-1-60566-954-0

Jennex, M. E. (2010). A specialized evaluation and comparison of

sample data mining software . In Ubiquitous developments in
knowledge management: Integrations and trends (pp. 300–318).
Hershey, PA: IGI Global. doi:10.4018/978-1-60566-954-0
Jennex, M. E. (2010). Using soft systems methodology to reveal
socio-technical barriers to knowledge sharing and management: A
case study from the UK national health service . In Ubiquitous
developments in knowledge management: Integrations and trends (pp.
215–235). Hershey, PA: IGI Global. doi:10.4018/978-1-60566-954-0

Jennex, M. E. (2011). Knowledge management success models . In

Schwartz, D., & Te’eni, D. (Eds.), Encyclopedia of knowledge
management (2nd ed.; pp. 763–771). Hershey, PA: IGI Global.
doi:10.4018/978-1-59904-931-1.ch074

Jennex, M. E., & Olfman, L. (2011). A model of knowledge

management success . In Jennex, M., & Smolnik, S. (Eds.), Strategies
for knowledge management success: Exploring organizational
efficacy (pp. 14–31). Hershey, PA: IGI Global. doi:10.4018/978-1-
60566-709-6.ch002

Jennex, M. E., Smolnik, S., & Croasdell, D. (2011). Towards a

consensus knowledge management success definition . In Jennex, M.,
& Smolnik, S. (Eds.), Strategies for knowledge management success:
Exploring organizational efficacy (pp. 1–13). Hershey, PA: IGI
Global. doi:10.4018/978-1-60566-709-6.ch001

Jewels, T. (2013). Teaching enterprise information systems in the

United Arab Emirates . In Albadri, F. (Ed.), Information systems
applications in the Arab education sector (pp. 322–337). Hershey, PA:
IGI Global. doi:10.4018/978-1-4666-1984-5.ch022

Jolly, R., & Wakeland, W. (2011). Using agent based simulation and
game theory analysis to study knowledge flow in organizations: The
KMscape . In Jennex, M. (Ed.), Global aspects and cultural
perspectives on knowledge management: Emerging dimensions (pp.
19–29). Hershey, PA: IGI Global. doi:10.4018/978-1-60960-555-
1.ch002
Joshi, S. (2014). Web 2.0 and its implications on globally competitive
business model . In Pańkowska, M. (Ed.), Frameworks of IT
prosumption for business development (pp. 86–101). Hershey, PA:
IGI Global. doi:10.4018/978-1-4666-4313-0.ch007

Judge, R. (2011). A simulation system for evaluating knowledge

management system (KMS) implementation strategies in small to
mid-size enterprises (SME) . In Jennex, M. (Ed.), Global aspects and
cultural perspectives on knowledge management: Emerging
dimensions (pp. 92–112). Hershey, PA: IGI Global. doi:10.4018/978-
1-60960-555-1.ch007

Kalid, K. S. (2011). Transfer knowledge using stories: A Malaysian

university case study . In Al-Shammari, M. (Ed.), Knowledge
management in emerging economies: Social, organizational and
cultural implementation (pp. 186–198). Hershey, PA: IGI Global.
doi:10.4018/978-1-61692-886-5.ch011

Kamau, C. (2010). Strategising impression management in

corporations: Cultural knowledge as capital . In Harorimana, D. (Ed.),
Cultural implications of knowledge sharing, management and transfer:
Identifying competitive advantage (pp. 60–83). Hershey, PA: IGI
Global. doi:10.4018/978-1-60566-790-4.ch004

Kamthan, P., & Fancott, T. (2011). A knowledge management model

for patterns . In Schwartz, D., & Te’eni, D. (Eds.), Encyclopedia of
knowledge management (2nd ed.; pp. 694–703). Hershey, PA: IGI
Global. doi:10.4018/978-1-59904-931-1.ch067

Kamthan, P., & Pai, H. (2011). Knowledge representation in pattern

management . In Schwartz, D., & Te’eni, D. (Eds.), Encyclopedia of
knowledge management (2nd ed.; pp. 893–904). Hershey, PA: IGI
Global. doi:10.4018/978-1-59904-931-1.ch085
Kane, G. C., Schwaig, K. S., & Storey, V. C. (2011). Information
privacy: Understanding how firms behave online . In Siau, K. (Ed.),
Theoretical and practical advances in information systems
development: Emerging trends and approaches (pp. 81–100). Hershey,
PA: IGI Global. doi:10.4018/978-1-60960-521-6.ch005

Kankanhalli, A., Tan, B. C., & Wei, K. (2011). Knowledge producers

and consumers . In Schwartz, D., & Te’eni, D. (Eds.), Encyclopedia
of knowledge management (2nd ed.; pp. 867–877). Hershey, PA: IGI
Global. doi:10.4018/978-1-59904-931-1.ch083

Karagiannis, D., Woitsch, R., & Hrgovcic, V. (2010). Industrialisation

of the knowledge work: The knowledge conveyer belt approach . In
O’Brien, E., Clifford, S., & Southern, M. (Eds.), Knowledge
management for process, organizational and marketing innovation:
Tools and methods (pp. 79–94). Hershey, PA: IGI Global.
doi:10.4018/978-1-61520-829-6.ch005

Karlsson, F., & Ågerfalk, P. J. (2011). Towards structured flexibility

in information systems development: Devising a method for method
configuration . In Siau, K. (Ed.), Theoretical and practical advances in
information systems development: Emerging trends and approaches
(pp. 214–238). Hershey, PA: IGI Global. doi:10.4018/978-1-60960-
521-6.ch010

Karna, A., Singh, R., & Verma, S. (2010). Knowledge management

for an effective sales and marketing function . In Russ, M. (Ed.),
Knowledge management strategies for business development (pp.
324–337). Hershey, PA: IGI Global. doi:10.4018/978-1-60566-348-
7.ch015

Kassim, A. M., & Cheah, Y. (2013). SEMblog: An ontology-based

semantic blogging tool for knowledge identification, organization, and
reuse . In Nazir Ahmad, M., Colomb, R., & Abdullah, M. (Eds.),
Ontology-based applications for enterprise systems and knowledge
management (pp. 210–223). Hershey, PA: IGI Global.
doi:10.4018/978-1-4666-1993-7.ch012

Kayakutlu, G. (2010). Knowledge worker profile: A framework to

clarify expectations . In Metaxiotis, K., Carrillo, F., & Yigitcanlar, T.
(Eds.), Knowledge-based development for cities and societies:
Integrated multi-level approaches (pp. 162–178). Hershey, PA: IGI
Global. doi:10.4018/978-1-61520-721-3.ch010

Kettunen, J., & Chaudhuri, M. R. (2011). Knowledge management to

promote organizational change in India . In Al-Shammari, M. (Ed.),
Knowledge management in emerging economies: Social,
organizational and cultural implementation (pp. 308–324). Hershey,
PA: IGI Global. doi:10.4018/978-1-61692-886-5.ch019

Khalil, O. E., & Seleim, A. (2012). Culture and knowledge transfer

capacity: A cross-national study . In Jennex, M. (Ed.), Conceptual
models and outcomes of advancing knowledge management: New
technologies (pp. 305–332). Hershey, PA: IGI Global.
doi:10.4018/978-1-4666-0035-5.ch016

Khasawneh, R., & Alazzam, A. (2014). Towards customer knowledge

management (CKM): Where knowledge and customer meet . In
Chilton, M., & Bloodgood, J. (Eds.), Knowledge management and
competitive advantage: Issues and potential solutions (pp. 109–121).
Hershey, PA: IGI Global. doi:10.4018/978-1-4666-4679-7.ch007

Kim, J. (2014). Big data sharing among academics . In Hu, W., &
Kaabouch, N. (Eds.), Big data management, technologies, and
applications (pp. 177–194). Hershey, PA: IGI Global.
doi:10.4018/978-1-4666-4699-5.ch008

Kim, S., Felan, J., & Kang, M. H. (2011). An ontological approach to

enterprise knowledge modeling in a shipping company. International
Journal of Knowledge Management , 7(4), 70–84.
doi:10.4018/jkm.2011100105

Kim, S., Felan, J., & Kang, M. H. (2013). An ontological approach to

enterprise knowledge modeling in a shipping company . In Jennex, M.
(Ed.), Dynamic models for knowledge-driven organizations (pp. 351–
363). Hershey, PA: IGI Global. doi:10.4018/978-1-4666-2485-
6.ch021

King, W. R. (2011). Knowledge transfer . In Schwartz, D., & Te’eni,

D. (Eds.), Encyclopedia of knowledge management (2nd ed.; pp. 967–
976). Hershey, PA: IGI Global. doi:10.4018/978-1-59904-931-
1.ch092

Kivijärvi, H., Piirainen, K., & Tuominen, M. (2010). Sustaining

organizational innovativeness: Advancing knowledge sharing during
the scenario process. International Journal of Knowledge Management
, 6(2), 22–39. doi:10.4018/jkm.2010040102

Kivijärvi, H., Piirainen, K., & Tuominen, M. (2012). Sustaining

organizational innovativeness: Advancing knowledge sharing during
the scenario process . In Jennex, M. (Ed.), Conceptual models and
outcomes of advancing knowledge management: New technologies
(pp. 99–117). Hershey, PA: IGI Global. doi:10.4018/978-1-4666-
0035-5.ch006

Knyazhansky, M., & Plotkin, T. (2012). Knowledge bases over

algebraic models: Some notes about informational equivalence.
International Journal of Knowledge Management , 8(1), 22–39.
doi:10.4018/jkm.2012010102

Kong, E. (2014). The role of social intelligence in acquiring external

knowledge for human capital development, organisational learning,
and innovation . In Chilton, M., & Bloodgood, J. (Eds.), Knowledge
management and competitive advantage: Issues and potential
solutions (pp. 53–70). Hershey, PA: IGI Global. doi:10.4018/978-1-
4666-4679-7.ch004

Kor, A., & Orange, G. (2011). A survey of epistemology and its

implications on an organisational information and knowledge
management model . In Eardley, A., & Uden, L. (Eds.), Innovative
knowledge management: Concepts for organizational creativity and
collaborative design (pp. 95–124). Hershey, PA: IGI Global.
doi:10.4018/978-1-60566-701-0.ch006

Kostrzewa, A., Laaksoharju, M., & Kavathatzopoulos, I. (2011).

Management of moral knowledge and ethical processes in
organizations . In Morais da Costa, G. (Ed.), Ethical issues and social
dilemmas in knowledge management: Organizational innovation (pp.
199–214). Hershey, PA: IGI Global. doi:10.4018/978-1-61520-873-
9.ch012

Kraaijenbrink, J., & Wijnhoven, F. (2011). External knowledge

integration . In Schwartz, D., & Te’eni, D. (Eds.), Encyclopedia of
knowledge management (2nd ed.; pp. 308–319). Hershey, PA: IGI
Global. doi:10.4018/978-1-59904-931-1.ch030

Kraft, T. A., & Steenkamp, A. L. (2012). A holistic approach for

understanding project management . In Stowell, F. (Ed.), Systems
approach applications for developments in information technology
(pp. 25–39). Hershey, PA: IGI Global. doi:10.4018/978-1-4666-1562-
5.ch003

Kulkarni, U., & Freeze, R. (2011). Measuring knowledge

management capabilities . In Schwartz, D., & Te’eni, D. (Eds.),
Encyclopedia of knowledge management (2nd ed.; pp. 1090–1100).
Hershey, PA: IGI Global. doi:10.4018/978-1-59904-931-1.ch104
Kumar, A. S., Alrabea, A., & Sekhar, P. C. (2013). Temporal
association rule mining in large databases . In Data mining: Concepts,
methodologies, tools, and applications (pp. 586–602). Hershey, PA:
IGI Global. doi:10.4018/978-1-4666-2455-9.ch029

Laihonen, H., & Koivuaho, M. (2011). Knowledge flow audit:

indentifying, measuring and managing knowledge asset dynamics . In
Vallejo-Alonso, B., Rodriguez-Castellanos, A., & Arregui-Ayastuy,
G. (Eds.), Identifying, measuring, and valuing knowledge-based
intangible assets: New perspectives (pp. 22–42). Hershey, PA: IGI
Global. doi:10.4018/978-1-60960-054-9.ch002

Land, F., Amjad, U., & Nolas, S. (2011). Knowledge management

processes . In Schwartz, D., & Te’eni, D. (Eds.), Encyclopedia of
knowledge management (2nd ed.; pp. 719–727). Hershey, PA: IGI
Global. doi:10.4018/978-1-59904-931-1.ch070

Lavanderos, L. P., & Fiol, E. S. (2011). Production cognitive capital

as a measurement of intellectual capital . In Vallejo-Alonso, B.,
Rodriguez-Castellanos, A., & Arregui-Ayastuy, G. (Eds.), Identifying,
measuring, and valuing knowledge-based intangible assets: New
perspectives (pp. 112–132). Hershey, PA: IGI Global.
doi:10.4018/978-1-60960-054-9.ch006

Lavoué, É., George, S., & Prévôt, P. (2011). A knowledge

management tool for the interconnection of communities of practice.
International Journal of Knowledge Management , 7(1), 55–76.
doi:10.4018/jkm.2011010104

Lee, H., Chan, K., & Tsui, E. (2013). Knowledge mining Wikipedia:
An ontological approach . In Yang, G. (Ed.), Multidisciplinary studies
in knowledge and systems science (pp. 52–62). Hershey, PA: IGI
Global. doi:10.4018/978-1-4666-3998-0.ch005
Leung, N. K. (2011). A re-distributed knowledge management
framework in help desk . In Schwartz, D., & Te’eni, D. (Eds.),
Encyclopedia of knowledge management (2nd ed.; pp. 1374–1381).
Hershey, PA: IGI Global. doi:10.4018/978-1-59904-931-1.ch131

Li, Y., Guo, H., & Wang, S. (2010). A multiple-bits watermark for
relational data . In Siau, K., & Erickson, J. (Eds.), Principle
advancements in database management technologies: New
applications and frameworks (pp. 1–22). Hershey, PA: IGI Global.
doi:10.4018/978-1-60566-904-5.ch001

Lin, C. Y. (2013). Intellectual capital explains a country’s resilience to

financial crisis: A resource-based view . In Ordóñez de Pablos, P.,
Tennyson, R., & Zhao, J. (Eds.), Intellectual capital strategy
management for knowledge-based organizations (pp. 52–75).
Hershey, PA: IGI Global. doi:10.4018/978-1-4666-3655-2.ch005

Lin, Y., & Dalkir, K. (2012). Factors affecting KM implementation in

the Chinese community . In Jennex, M. (Ed.), Conceptual models and
outcomes of advancing knowledge management: New technologies
(pp. 1–23). Hershey, PA: IGI Global. doi:10.4018/978-1-4666-0035-
5.ch001

Lindsey, K. L. (2011). Barriers to knowledge sharing . In Schwartz,

D., & Te’eni, D. (Eds.), Encyclopedia of knowledge management
(2nd ed.; pp. 49–61). Hershey, PA: IGI Global. doi:10.4018/978-1-
59904-931-1.ch006

Liu, K., Tan, H. B., & Chen, X. (2013). Aiding maintenance of

database applications through extracting attribute dependency graph.
Journal of Database Management , 24(1), 20–35.
doi:10.4018/jdm.2013010102

Liu, K., Tan, H. B., & Chen, X. (2013). Automated insertion of

exception handling for key and referential constraints. Journal of
Database Management , 24(1), 1–19. doi:10.4018/jdm.2013010101

Locuratolo, E., & Palomäki, J. (2013). Ontology for database

preservation . In Nazir Ahmad, M., Colomb, R., & Abdullah, M.
(Eds.), Ontology-based applications for enterprise systems and
knowledge management (pp. 141–157). Hershey, PA: IGI Global.
doi:10.4018/978-1-4666-1993-7.ch008

López-Nicolás, C., & Meroño-Cerdán, Á. L. (2010). A model for

knowledge management and intellectual capital audits . In Russ, M.
(Ed.), Knowledge management strategies for business development
(pp. 115–131). Hershey, PA: IGI Global. doi:10.4018/978-1-60566-
348-7.ch006

Luck, D. (2010). The implications of the development and

implementation of CRM for knowledge management . In Russ, M.
(Ed.), Knowledge management strategies for business development
(pp. 338–352). Hershey, PA: IGI Global. doi:10.4018/978-1-60566-
348-7.ch016

Lukovic, I., Ivancevic, V., Celikovic, M., & Aleksic, S. (2014). DSLs
in action with model based approaches to information system
development . In Software design and development: Concepts,
methodologies, tools, and applications (pp. 596–626). Hershey, PA:
IGI Global. doi:10.4018/978-1-4666-4301-7.ch029

Luna-Reyes, L. F., & Gil-Garcia, J. R. (2012). Government and inter-

organizational collaboration as strategies for administrative reform in
Mexico . In Papadopoulos, T., & Kanellis, P. (Eds.), Public sector
reform using information technologies: Transforming policy into
practice (pp. 79–101). Hershey, PA: IGI Global. doi:10.4018/978-1-
60960-839-2.ch005
Lungu, C. I., Caraiani, C., & Dascalu, C. (2013). Sustainable
intellectual capital: The inference of corporate social responsibility
within intellectual capital . In Ordóñez de Pablos, P., Tennyson, R., &
Zhao, J. (Eds.), Intellectual capital strategy management for
knowledge-based organizations (pp. 156–173). Hershey, PA: IGI
Global. doi:10.4018/978-1-4666-3655-2.ch009

Ma, Z. M. (2011). Engineering design knowledge management . In

Schwartz, D., & Te’eni, D. (Eds.), Encyclopedia of knowledge
management (2nd ed.; pp. 263–269). Hershey, PA: IGI Global.
doi:10.4018/978-1-59904-931-1.ch025

Maier, R., & Hadrich, T. (2011). Knowledge management systems .

In Schwartz, D., & Te’eni, D. (Eds.), Encyclopedia of knowledge
management (2nd ed.; pp. 779–790). Hershey, PA: IGI Global.
doi:10.4018/978-1-59904-931-1.ch076

Maria, E. D., & Micelli, S. (2010). SMEs and competitive advantage:

A mix of innovation, marketing and ICT—The case of “made in Italy”
. In Russ, M. (Ed.), Knowledge management strategies for business
development (pp. 310–323). Hershey, PA: IGI Global.
doi:10.4018/978-1-60566-348-7.ch014

Mariano, S., & Simionato, N. (2010). Where are we looking? A

practical approach to managing knowledge captured from eye-
tracking experiments: The experience of gulf air . In O’Brien, E.,
Clifford, S., & Southern, M. (Eds.), Knowledge management for
process, organizational and marketing innovation: Tools and methods
(pp. 216–227). Hershey, PA: IGI Global. doi:10.4018/978-1-61520-
829-6.ch013

Marques, M. B. (2014). The value of information and information

services in knowledge society . In Jamil, G., Malheiro, A., & Ribeiro,
F. (Eds.), Rethinking the conceptual base for new practical
applications in information value and quality (pp. 134–161). Hershey,
PA: IGI Global. doi:10.4018/978-1-4666-4562-2.ch007

Masrom, M., Mahmood, N. H., & Al-Araimi, A. A. (2014). Exploring

knowledge types and knowledge protection in organizations . In
Chilton, M., & Bloodgood, J. (Eds.), Knowledge management and
competitive advantage: Issues and potential solutions (pp. 271–280).
Hershey, PA: IGI Global. doi:10.4018/978-1-4666-4679-7.ch016

Masterson, F. (2013). Knowledge management in practice: Using

wikis to facilitate project-based learning . In Buckley, S., &
Jakovljevic, M. (Eds.), Knowledge management innovations for
interdisciplinary education: Organizational applications (pp. 385–
401). Hershey, PA: IGI Global. doi:10.4018/978-1-4666-1969-
2.ch019

Mattmann, C. A., Hart, A., Cinquini, L., Lazio, J., Khudikyan, S., &
Jones, D. … Robnett, J. (2014). Scalable data mining, archiving, and
big data management for the next generation astronomical telescopes.
In W. Hu, & N. Kaabouch (Eds.), Big data management,
technologies, and applications (pp. 196-221). Hershey, PA: IGI
Global. doi:doi:10.4018/978-1-4666-4699-5.ch009

Maule, R. W. (2011). Military knowledge management . In Schwartz,

D., & Te’eni, D. (Eds.), Encyclopedia of knowledge management
(2nd ed.; pp. 1125–1135). Hershey, PA: IGI Global. doi:10.4018/978-
1-59904-931-1.ch108

Mavridis, I. (2011). Deploying privacy improved RBAC in web

information systems. International Journal of Information
Technologies and Systems Approach , 4(2), 70–87.
doi:10.4018/jitsa.2011070105

Mavridis, I. (2012). Deploying privacy improved RBAC in web

information systems . In Stowell, F. (Ed.), Systems approach
applications for developments in information technology (pp. 298–
315). Hershey, PA: IGI Global. doi:10.4018/978-1-4666-1562-
5.ch020

McLaughlin, S. (2011). Assessing the impact of knowledge transfer

mechanisms on supply chain performance . In Jennex, M. (Ed.),
Global aspects and cultural perspectives on knowledge management:
Emerging dimensions (pp. 157–171). Hershey, PA: IGI Global.
doi:10.4018/978-1-60960-555-1.ch011

Medina, J. M., & Spinola, M. D. (2011). Understanding the behavior

of knowledge management pathways: The case of small
manufacturers of footwear in Peru and Brazil . In Al-Shammari, M.
(Ed.), Knowledge management in emerging economies: Social,
organizational and cultural implementation (pp. 261–271). Hershey,
PA: IGI Global. doi:10.4018/978-1-61692-886-5.ch016

Meloche, J. A., Hasan, H., Willis, D., Pfaff, C. C., & Qi, Y. (2011).
Cocreating corporate knowledge with a wiki . In Jennex, M. (Ed.),
Global aspects and cultural perspectives on knowledge management:
Emerging dimensions (pp. 126–143). Hershey, PA: IGI Global.
doi:10.4018/978-1-60960-555-1.ch009

Melzer, S. (2013). On the relationship between ontology-based and

holistic representations in a knowledge management system . In Nazir
Ahmad, M., Colomb, R., & Abdullah, M. (Eds.), Ontology-based
applications for enterprise systems and knowledge management (pp.
292–323). Hershey, PA: IGI Global. doi:10.4018/978-1-4666-1993-
7.ch017

Mendes, E., & Baker, S. (2013). Using knowledge management and

aggregation techniques to improve web effort estimation . In Saeed,
S., & Alsmadi, I. (Eds.), Knowledge-based processes in software
development (pp. 64–85). Hershey, PA: IGI Global. doi:10.4018/978-
1-4666-4229-4.ch005

Metaxiotis, K. (2011). Healthcare knowledge management . In

Schwartz, D., & Te’eni, D. (Eds.), Encyclopedia of knowledge
management (2nd ed.; pp. 366–375). Hershey, PA: IGI Global.
doi:10.4018/978-1-59904-931-1.ch035

Mikolajuk, Z. (2013). Community-based development of knowledge

products . In Jennex, M. (Ed.), Dynamic models for knowledge-driven
organizations (pp. 268–281). Hershey, PA: IGI Global.
doi:10.4018/978-1-4666-2485-6.ch016

Mischo, W. H., Schlembach, M. C., Bishoff, J., & German, E. M.

(2012). User search activities within an academic library gateway:
Implications for web-scale discovery systems . In Popp, M., & Dallis,
D. (Eds.), Planning and implementing resource discovery tools in
academic libraries (pp. 153–173). Hershey, PA: IGI Global.
doi:10.4018/978-1-4666-1821-3.ch010

Mishra, B., & Shukla, K. K. (2014). Data mining techniques for

software quality prediction . In Software design and development:
Concepts, methodologies, tools, and applications (pp. 401–428).
Hershey, PA: IGI Global. doi:10.4018/978-1-4666-4301-7.ch021

Moffett, S., Walker, T., & McAdam, R. (2014). Best value and
performance management inspired change within UK councils: A
knowledge management perspective . In Al-Bastaki, Y., & Shajera, A.
(Eds.), Building a competitive public sector with knowledge
management strategy (pp. 199–226). Hershey, PA: IGI Global.
doi:10.4018/978-1-4666-4434-2.ch009

Mueller, C. E., & Bradley, K. D. (2011). Utilizing the Rasch model to

develop and evaluate items for the tacit knowledge inventory for
superintendents (TKIS) . In Jennex, M. (Ed.), Global aspects and
cultural perspectives on knowledge management: Emerging
dimensions (pp. 264–284). Hershey, PA: IGI Global.
doi:10.4018/978-1-60960-555-1.ch017

Muhammed, S., Doll, W. J., & Deng, X. (2011). Impact of knowledge

management practices on task knowledge: An individual level study.
International Journal of Knowledge Management , 7(4), 1–21.
doi:10.4018/jkm.2011100101

Muhammed, S., Doll, W. J., & Deng, X. (2011). Measuring

knowledge management outcomes at the individual level: Towards a
tool for research on organizational culture . In Jennex, M. (Ed.),
Global aspects and cultural perspectives on knowledge management:
Emerging dimensions (pp. 1–18). Hershey, PA: IGI Global.
doi:10.4018/978-1-60960-555-1.ch001

Muhammed, S., Doll, W. J., & Deng, X. (2013). Impact of knowledge

management practices on task knowledge: An individual level study .
In Jennex, M. (Ed.), Dynamic models for knowledge-driven
organizations (pp. 282–301). Hershey, PA: IGI Global.
doi:10.4018/978-1-4666-2485-6.ch017

Murata, K. (2011). Knowledge creation and sharing in Japanese

organisations: A socio-cultural perspective on ba . In Morais da Costa,
G. (Ed.), Ethical issues and social dilemmas in knowledge
management: Organizational innovation (pp. 1–16). Hershey, PA: IGI
Global. doi:10.4018/978-1-61520-873-9.ch001

Murphy, P. (2013). Systems of communication: Information,

explanation, and imagination . In Yang, G. (Ed.), Multidisciplinary
studies in knowledge and systems science (pp. 63–78). Hershey, PA:
IGI Global. doi:10.4018/978-1-4666-3998-0.ch006
Nach, H. (2013). Structuring knowledge for enterprise resource
planning implementation through an ontology . In Nazir Ahmad, M.,
Colomb, R., & Abdullah, M. (Eds.), Ontology-based applications for
enterprise systems and knowledge management (pp. 25–42). Hershey,
PA: IGI Global. doi:10.4018/978-1-4666-1993-7.ch002

Nah, F. F., Hong, W., Chen, L., & Lee, H. (2010). Information search
patterns in e-commerce product comparison services. Journal of
Database Management , 21(2), 26–40. doi:10.4018/jdm.2010040102

Nah, F. F., Hong, W., Chen, L., & Lee, H. (2012). Information search
patterns in e-commerce product comparison services . In Siau, K.
(Ed.), Cross-disciplinary models and applications of database
management: Advancing approaches (pp. 131–145). Hershey, PA: IGI
Global. doi:10.4018/978-1-61350-471-0.ch006

Natarajan, R., & Shekar, B. (2011). Knowledge patterns in databases .

In Schwartz, D., & Te’eni, D. (Eds.), Encyclopedia of knowledge
management (2nd ed.; pp. 842–852). Hershey, PA: IGI Global.
doi:10.4018/978-1-59904-931-1.ch081

Nelson, R. E., & Hsu, H. S. (2011). A social network perspective on

knowledge management . In Schwartz, D., & Te’eni, D. (Eds.),
Encyclopedia of knowledge management (2nd ed.; pp. 1470–1478).
Hershey, PA: IGI Global. doi:10.4018/978-1-59904-931-1.ch140

Neto, R. C., & Souza, R. R. (2010). Knowledge management as an

organizational process: From a theoretical framework to
implementation guidelines . In O’Brien, E., Clifford, S., & Southern,
M. (Eds.), Knowledge management for process, organizational and
marketing innovation: Tools and methods (pp. 16–35). Hershey, PA:
IGI Global. doi:10.4018/978-1-61520-829-6.ch002

Newell, S. (2011). Understanding innovation processes . In Schwartz,

D., & Te’eni, D. (Eds.), Encyclopedia of knowledge management
(2nd ed.; pp. 1525–1535). Hershey, PA: IGI Global. doi:10.4018/978-
1-59904-931-1.ch145

Nikabadi, M. S., & Zamanloo, S. (2012). A multidimensional

structure for describing the influence of supply chain strategies,
business strategies, and knowledge management strategies on
knowledge sharing in supply chain. International Journal of
Knowledge Management , 8(4), 50–70. doi:10.4018/jkm.2012100103

Nissen, M. (2014). Cyberspace and cloud knowledge . In Harnessing

dynamic knowledge principles in the technology-driven world (pp.
193–204). Hershey, PA: IGI Global. doi:10.4018/978-1-4666-4727-
5.ch012

Nissen, M. (2014). Social media knowledge . In Harnessing dynamic

knowledge principles in the technology-driven world (pp. 219–227).
Hershey, PA: IGI Global. doi:10.4018/978-1-4666-4727-5.ch014

Nissen, M. E. (2014). Harnessing knowledge power for competitive

advantage . In Chilton, M., & Bloodgood, J. (Eds.), Knowledge
management and competitive advantage: Issues and potential
solutions (pp. 20–34). Hershey, PA: IGI Global. doi:10.4018/978-1-
4666-4679-7.ch002

Nissen, M. E., & Levitt, R. E. (2011). Knowledge management

research through computational experimentation . In Schwartz, D., &
Te’eni, D. (Eds.), Encyclopedia of knowledge management (2nd ed.;
pp. 728–737). Hershey, PA: IGI Global. doi:10.4018/978-1-59904-
931-1.ch071

Nisula, A. (2014). Developing organizational renewal capability in the

municipal (city) organization . In Al-Bastaki, Y., & Shajera, A. (Eds.),
Building a competitive public sector with knowledge management
strategy (pp. 151–172). Hershey, PA: IGI Global. doi:10.4018/978-1-
4666-4434-2.ch007

Niu, B., Martin, P., & Powley, W. (2011). Towards autonomic

workload management in DBMSs . In Siau, K. (Ed.), Theoretical and
practical advances in information systems development: Emerging
trends and approaches (pp. 154–173). Hershey, PA: IGI Global.
doi:10.4018/978-1-60960-521-6.ch008

Nobre, F. S., & Walker, D. S. (2011). A dynamic ability-based view

of the organization. International Journal of Knowledge Management ,
7(2), 86–101. doi:10.4018/jkm.2011040105

O’Brien, J. (2014). Lessons from the private sector: A framework to

be adopted in the public sector . In Al-Bastaki, Y., & Shajera, A.
(Eds.), Building a competitive public sector with knowledge
management strategy (pp. 173–198). Hershey, PA: IGI Global.
doi:10.4018/978-1-4666-4434-2.ch008

Omari, A. (2013). Supporting companies management and improving

their productivity through mining customers transactions . In Data
mining: Concepts, methodologies, tools, and applications (pp. 1519–
1533). Hershey, PA: IGI Global. doi:10.4018/978-1-4666-2455-
9.ch079

Onwubiko, C. (2014). Modelling situation awareness information and

system requirements for the mission using goal-oriented task analysis
approach . In Software design and development: Concepts,
methodologies, tools, and applications (pp. 460–478). Hershey, PA:
IGI Global. doi:10.4018/978-1-4666-4301-7.ch023

Orth, A., Smolnik, S., & Jennex, M. E. (2011). The relevance of

integration for knowledge management success: Towards conceptual
and empirical evidence . In Jennex, M., & Smolnik, S. (Eds.),
Strategies for knowledge management success: Exploring
organizational efficacy (pp. 238–261). Hershey, PA: IGI Global.
doi:10.4018/978-1-60566-709-6.ch013

Othman, A. K., & Abdullah, H. S. (2011). The influence of emotional

intelligence on tacit knowledge sharing in service organizations . In
Al-Shammari, M. (Ed.), Knowledge management in emerging
economies: Social, organizational and cultural implementation (pp.
171–185). Hershey, PA: IGI Global. doi:10.4018/978-1-61692-886-
5.ch010

Pagallo, U. (2011). The trouble with digital copies: A short KM

phenomenology . In Morais da Costa, G. (Ed.), Ethical issues and
social dilemmas in knowledge management: Organizational
innovation (pp. 97–112). Hershey, PA: IGI Global. doi:10.4018/978-
1-61520-873-9.ch007

Palte, R., Hertlein, M., Smolnik, S., & Riempp, G. (2013). The effects
of a KM strategy on KM performance in professional services firms .
In Jennex, M. (Ed.), Dynamic models for knowledge-driven
organizations (pp. 16–35). Hershey, PA: IGI Global. doi:10.4018/978-
1-4666-2485-6.ch002

Pańkowska, M. (2014). Information technology prosumption

acceptance by business information system consultants . In
Pańkowska, M. (Ed.), Frameworks of IT prosumption for business
development (pp. 119–141). Hershey, PA: IGI Global.
doi:10.4018/978-1-4666-4313-0.ch009

Pankowski, T. (2011). Pattern-based schema mapping and query

answering in peer-to-peer XML data integration system . In Yan, L.,
& Ma, Z. (Eds.), Advanced database query systems: Techniques,
applications and technologies (pp. 221–246). Hershey, PA: IGI
Global. doi:10.4018/978-1-60960-475-2.ch009
Papoutsakis, H. (2010). New product development based on
knowledge creation and technology education . In O’Brien, E.,
Clifford, S., & Southern, M. (Eds.), Knowledge management for
process, organizational and marketing innovation: Tools and methods
(pp. 148–163). Hershey, PA: IGI Global. doi:10.4018/978-1-61520-
829-6.ch009

Paquette, S. (2011). Applying knowledge management in the

environmental and climate change sciences . In Schwartz, D., &
Te’eni, D. (Eds.), Encyclopedia of knowledge management (2nd ed.;
pp. 20–26). Hershey, PA: IGI Global. doi:10.4018/978-1-59904-931-
1.ch003

Paquette, S. (2011). Customer knowledge management . In Schwartz,

D., & Te’eni, D. (Eds.), Encyclopedia of knowledge management
(2nd ed.; pp. 175–184). Hershey, PA: IGI Global. doi:10.4018/978-1-
59904-931-1.ch017

Parker, K. R., & Nitse, P. S. (2011). Competitive intelligence

gathering . In Schwartz, D., & Te’eni, D. (Eds.), Encyclopedia of
knowledge management (2nd ed.; pp. 103–111). Hershey, PA: IGI
Global. doi:10.4018/978-1-59904-931-1.ch011

Páscoa, C., & Tribolet, J. (2014). Maintaining organizational viability

and performance: The organizational configuration map . In Jamil, G.,
Malheiro, A., & Ribeiro, F. (Eds.), Rethinking the conceptual base for
new practical applications in information value and quality (pp. 266–
283). Hershey, PA: IGI Global. doi:10.4018/978-1-4666-4562-
2.ch012

Paukert, M., Niederée, C., & Hemmje, M. (2011). Knowledge in

innovation processes . In Schwartz, D., & Te’eni, D. (Eds.),
Encyclopedia of knowledge management (2nd ed.; pp. 570–580).
Hershey, PA: IGI Global. doi:10.4018/978-1-59904-931-1.ch055
Pawlak, P. (2011). Global “knowledge management” in humanist
perspective. In G. Morais da Costa (Ed.), Ethical issues and social
dilemmas in knowledge management: Organizational innovation (pp.
45-62). Hershey, PA: IGI Global. doi:doi:10.4018/978-1-61520-873-
9.ch004

Perry, M. (2013). Strategic knowledge management: A university

application . In Buckley, S., & Jakovljevic, M. (Eds.), Knowledge
management innovations for interdisciplinary education:
Organizational applications (pp. 132–144). Hershey, PA: IGI Global.
doi:10.4018/978-1-4666-1969-2.ch007

Pessoa, C. R., Silva, U. P., & Cruz, C. H. (2014). Information

management in industrial areas: A knowledge management view . In
Jamil, G., Malheiro, A., & Ribeiro, F. (Eds.), Rethinking the
conceptual base for new practical applications in information value
and quality (pp. 378–395). Hershey, PA: IGI Global.
doi:10.4018/978-1-4666-4562-2.ch017

Peter, H., & Greenidge, C. (2011). An ontology-based extraction

framework for a semantic web application. International Journal of
Knowledge-Based Organizations , 1(3), 56–71.
doi:10.4018/ijkbo.2011070104

Peter, H., & Greenidge, C. (2013). An ontology-based extraction

framework for a semantic web application . In Wang, J. (Ed.),
Intelligence methods and systems advancements for knowledge-based
business (pp. 231–246). Hershey, PA: IGI Global. doi:10.4018/978-1-
4666-1873-2.ch013

Pham, Q. T., & Hara, Y. (2011). KM approach for improving the

labor productivity of Vietnamese enterprise. International Journal of
Knowledge Management , 7(3), 27–42. doi:10.4018/jkm.2011070103
Pham, Q. T., & Hara, Y. (2013). KM approach for improving the
labor productivity of Vietnamese enterprise . In Jennex, M. (Ed.),
Dynamic models for knowledge-driven organizations (pp. 206–219).
Hershey, PA: IGI Global. doi:10.4018/978-1-4666-2485-6.ch012

Philpott, E., & Beaumont-Kerridge, J. (2010). Overcoming reticence

to aid knowledge creation between universities and business: A case
reviewed . In Harorimana, D. (Ed.), Cultural implications of
knowledge sharing, management and transfer: Identifying competitive
advantage (pp. 355–368). Hershey, PA: IGI Global. doi:10.4018/978-
1-60566-790-4.ch016

Pike, S., & Roos, G. (2011). Measuring and valuing knowledge-based

intangible assets: Real business uses . In Vallejo-Alonso, B.,
Rodriguez-Castellanos, A., & Arregui-Ayastuy, G. (Eds.), Identifying,
measuring, and valuing knowledge-based intangible assets: New
perspectives (pp. 268–293). Hershey, PA: IGI Global.
doi:10.4018/978-1-60960-054-9.ch013

Pineda, J. L., Zapata, L. E., & Ramírez, J. (2010). Strengthening

knowledge transfer between the university and enterprise: A
conceptual model for collaboration . In Harorimana, D. (Ed.), Cultural
implications of knowledge sharing, management and transfer:
Identifying competitive advantage (pp. 134–151). Hershey, PA: IGI
Global. doi:10.4018/978-1-60566-790-4.ch007

Platonov, V., & Bergman, J. (2013). Cross-border cooperative

network in the perspective of innovation dynamics . In Wang, J. (Ed.),
Intelligence methods and systems advancements for knowledge-based
business (pp. 150–169). Hershey, PA: IGI Global. doi:10.4018/978-1-
4666-1873-2.ch009

Poels, G. (2013). Understanding business domain models: The effect

of recognizing resource-event-agent conceptual modeling structures .
In Siau, K. (Ed.), Innovations in database design, web applications,
and information systems management (pp. 72–106). Hershey, PA: IGI
Global. doi:10.4018/978-1-4666-2044-5.ch004

Poels, G., Decreus, K., Roelens, B., & Snoeck, M. (2013).

Investigating goal-oriented requirements engineering for business
processes. Journal of Database Management , 24(2), 35–71.
doi:10.4018/jdm.2013040103

Pomares-Quimbaya, A., & Torres-Moreno, M. E. (2013). Knowledge

management processes supported by ontology technologies . In Nazir
Ahmad, M., Colomb, R., & Abdullah, M. (Eds.), Ontology-based
applications for enterprise systems and knowledge management (pp.
125–140). Hershey, PA: IGI Global. doi:10.4018/978-1-4666-1993-
7.ch007

Ponis, S. T., Vagenas, G., & Koronis, E. (2010). Exploring the

knowledge management landscape: A critical review of existing
knowledge management frameworks . In Harorimana, D. (Ed.),
Cultural implications of knowledge sharing, management and transfer:
Identifying competitive advantage (pp. 1–25). Hershey, PA: IGI
Global. doi:10.4018/978-1-60566-790-4.ch001

Powers, S. M., & Salmon, C. (2010). Management of learning space .

In Wu, D. (Ed.), Temporal structures in individual time management:
Practices to enhance calendar tool design (pp. 210–219). Hershey, PA:
IGI Global. doi:10.4018/978-1-60566-776-8.ch015

Pretorius, A. B., & Coetzee, F. P. (2011). Model of a knowledge

management support system for choosing intellectual capital
assessment methods . In Vallejo-Alonso, B., Rodriguez-Castellanos,
A., & Arregui-Ayastuy, G. (Eds.), Identifying, measuring, and valuing
knowledge-based intangible assets: New perspectives (pp. 336–359).
Hershey, PA: IGI Global. doi:10.4018/978-1-60960-054-9.ch016
Pullinger, D. (2011). Mobilizing knowledge in the UK public sector:
Current issues and discourse . In Morais da Costa, G. (Ed.), Ethical
issues and social dilemmas in knowledge management:
Organizational innovation (pp. 232–249). Hershey, PA: IGI Global.
doi:10.4018/978-1-61520-873-9.ch014

Rabaey, M. (2013). Complex adaptive systems thinking approach for

intelligence base in support of intellectual capital management . In
Ordóñez de Pablos, P., Tennyson, R., & Zhao, J. (Eds.), Intellectual
capital strategy management for knowledge-based organizations (pp.
122–141). Hershey, PA: IGI Global. doi:10.4018/978-1-4666-3655-
2.ch007

Rabaey, M., & Mercken, R. (2013). Framework of knowledge and

intelligence base: From intelligence to service . In Data mining:
Concepts, methodologies, tools, and applications (pp. 474–502).
Hershey, PA: IGI Global. doi:10.4018/978-1-4666-2455-9.ch023

Radziwill, N. M., & DuPlain, R. F. (2010). Quality and continuous

improvement in knowledge management . In Russ, M. (Ed.),
Knowledge management strategies for business development (pp.
353–363). Hershey, PA: IGI Global. doi:10.4018/978-1-60566-348-
7.ch017

Rahman, B. A., Saad, N. M., & Harun, M. S. (2010). Knowledge

management orientation and business performance: The Malaysian
manufacturing and service industries perspective . In Metaxiotis, K.,
Carrillo, F., & Yigitcanlar, T. (Eds.), Knowledge-based development
for cities and societies: Integrated multi-level approaches (pp. 315–
328). Hershey, PA: IGI Global. doi:10.4018/978-1-61520-721-
3.ch019

Randles, T. J., Blades, C. D., & Fadlalla, A. (2012). The knowledge

spectrum. International Journal of Knowledge Management , 8(2),
65–78. doi:10.4018/jkm.2012040104

Real, J. C., Leal, A., & Roldan, J. L. (2011). Measuring organizational

learning as a multidimensional construct . In Schwartz, D., & Te’eni,
D. (Eds.), Encyclopedia of knowledge management (2nd ed.; pp.
1101–1109). Hershey, PA: IGI Global. doi:10.4018/978-1-59904-931-
1.ch105

Rech, J., & Bogner, C. (2010). Qualitative analysis of semantically

enabled knowledge management systems in agile software
engineering. International Journal of Knowledge Management , 6(2),
66–85. doi:10.4018/jkm.2010040104

Rech, J., & Bogner, C. (2012). Qualitative analysis of semantically

enabled knowledge management systems in agile software
engineering . In Jennex, M. (Ed.), Conceptual models and outcomes
of advancing knowledge management: New technologies (pp. 144–
164). Hershey, PA: IGI Global. doi:10.4018/978-1-4666-0035-
5.ch008

Reis, R. S., & Curzi, Y. (2011). Knowledge integration in the creative

process of globally distributed teams . In Al-Shammari, M. (Ed.),
Knowledge management in emerging economies: Social,
organizational and cultural implementation (pp. 47–65). Hershey, PA:
IGI Global. doi:10.4018/978-1-61692-886-5.ch003

Remli, M. A., & Deris, S. (2013). An approach for biological data

integration and knowledge retrieval based on ontology, semantic web
services composition, and AI planning . In Nazir Ahmad, M., Colomb,
R., & Abdullah, M. (Eds.), Ontology-based applications for enterprise
systems and knowledge management (pp. 324–342). Hershey, PA:
IGI Global. doi:10.4018/978-1-4666-1993-7.ch018

Reychav, I., Stein, E. W., Weisberg, J., & Glezer, C. (2012). The role
of knowledge sharing in raising the task innovativeness of systems
analysts. International Journal of Knowledge Management , 8(2), 1–
22. doi:10.4018/jkm.2012040101

Reychav, I., & Weisberg, J. (2011). Human capital in knowledge

creation, management, and utilization . In Schwartz, D., & Te’eni, D.
(Eds.), Encyclopedia of knowledge management (2nd ed.; pp. 389–
401). Hershey, PA: IGI Global. doi:10.4018/978-1-59904-931-
1.ch037

Rhoads, E., O’Sullivan, K. J., & Stankosky, M. (2011). An evaluation

of factors that influence the success of knowledge management
practices in US federal agencies . In Jennex, M., & Smolnik, S. (Eds.),
Strategies for knowledge management success: Exploring
organizational efficacy (pp. 74–90). Hershey, PA: IGI Global.
doi:10.4018/978-1-60566-709-6.ch005

Ribière, V. M. (2011). The effect of organizational trust on the

success of codification and personalization KM approaches . In
Jennex, M., & Smolnik, S. (Eds.), Strategies for knowledge
management success: Exploring organizational efficacy (pp. 192–
212). Hershey, PA: IGI Global. doi:10.4018/978-1-60566-709-
6.ch011

Ribière, V. M., & Román, J. A. (2011). Knowledge flow . In

Schwartz, D., & Te’eni, D. (Eds.), Encyclopedia of knowledge
management (2nd ed.; pp. 549–559). Hershey, PA: IGI Global.
doi:10.4018/978-1-59904-931-1.ch053

Ricceri, F., Guthrie, J., & Coyte, R. (2010). The management of

knowledge resources within private organisations: Some European
“better practice” illustrations . In O’Brien, E., Clifford, S., &
Southern, M. (Eds.), Knowledge management for process,
organizational and marketing innovation: Tools and methods (pp. 36–
61). Hershey, PA: IGI Global. doi:10.4018/978-1-61520-829-6.ch003

Riss, U. V. (2011). Pattern-based task management as means of

organizational knowledge maturing. International Journal of
Knowledge-Based Organizations , 1(1), 20–41.
doi:10.4018/ijkbo.2011010102

Riss, U. V. (2013). Pattern-based task management as means of

organizational knowledge maturing . In Wang, J. (Ed.), Intelligence
methods and systems advancements for knowledge-based business
(pp. 1–23). Hershey, PA: IGI Global. doi:10.4018/978-1-4666-1873-
2.ch001

Roos, G. (2013). The role of intellectual capital in business model

innovation: An empirical study . In Ordóñez de Pablos, P., Tennyson,
R., & Zhao, J. (Eds.), Intellectual capital strategy management for
knowledge-based organizations (pp. 76–121). Hershey, PA: IGI
Global. doi:10.4018/978-1-4666-3655-2.ch006

Rothberg, H. N., & Klingenberg, B. (2010). Learning before doing: A

theoretical perspective and practical lessons from a failed cross-border
knowledge transfer initiative . In Harorimana, D. (Ed.), Cultural
implications of knowledge sharing, management and transfer:
Identifying competitive advantage (pp. 277–294). Hershey, PA: IGI
Global. doi:10.4018/978-1-60566-790-4.ch013

Ruano-Mayoral, M., Colomo-Palacios, R., García-Crespo, Á., &

Gómez-Berbís, J. M. (2012). Software project managers under the
team software process: A study of competences based on literature . In
Wang, J. (Ed.), Project management techniques and innovations in
information technology (pp. 115–126). Hershey, PA: IGI Global.
doi:10.4018/978-1-4666-0930-3.ch007

Russell, S. (2010). Knowledge management and project management

in 3D: A virtual world extension . In O’Brien, E., Clifford, S., &
Southern, M. (Eds.), Knowledge management for process,
organizational and marketing innovation: Tools and methods (pp. 62–
78). Hershey, PA: IGI Global. doi:10.4018/978-1-61520-829-6.ch004

Ryan, G., & Shinnick, E. (2011). Knowledge and intellectual property

rights: An economics perspective . In Schwartz, D., & Te’eni, D.
(Eds.), Encyclopedia of knowledge management (2nd ed.; pp. 489–
496). Hershey, PA: IGI Global. doi:10.4018/978-1-59904-931-
1.ch047

Sabetzadeh, F., & Tsui, E. (2013). Delivering knowledge services in

the cloud . In Yang, G. (Ed.), Multidisciplinary studies in knowledge
and systems science (pp. 247–254). Hershey, PA: IGI Global.
doi:10.4018/978-1-4666-3998-0.ch017

Sáenz, J., & Aramburu, N. (2011). Organizational conditions as

catalysts for successful people-focused knowledge sharing initiatives:
An empirical study. International Journal of Knowledge-Based
Organizations , 1(2), 39–56. doi:10.4018/ijkbo.2011040103

Sáenz, J., & Aramburu, N. (2013). Organizational conditions as

catalysts for successful people-focused knowledge sharing initiatives:
An empirical study . In Wang, J. (Ed.), Intelligence methods and
systems advancements for knowledge-based business (pp. 263–280).
Hershey, PA: IGI Global. doi:10.4018/978-1-4666-1873-2.ch015

Sakr, S., & Al-Naymat, G. (2011). Relational techniques for storing

and querying RDF data: An overview . In Yan, L., & Ma, Z. (Eds.),
Advanced database query systems: Techniques, applications and
technologies (pp. 269–285). Hershey, PA: IGI Global.
doi:10.4018/978-1-60960-475-2.ch011

Salem, P. J. (2013). The use of mixed methods in organizational

communication research . In Bocarnea, M., Reynolds, R., & Baker, J.
(Eds.), Online instruments, data collection, and electronic
measurements: Organizational advancements (pp. 24–39). Hershey,
PA: IGI Global. doi:10.4018/978-1-4666-2172-5.ch002

Salisbury, M. (2011). A framework for managing the life cycle of

knowledge in global organizations . In Jennex, M. (Ed.), Global
aspects and cultural perspectives on knowledge management:
Emerging dimensions (pp. 64–80). Hershey, PA: IGI Global.
doi:10.4018/978-1-60960-555-1.ch005

Salleh, K. (2014). Drivers, benefits, and challenges of knowledge

management in electronic government: Preliminary examination . In
Al-Bastaki, Y., & Shajera, A. (Eds.), Building a competitive public
sector with knowledge management strategy (pp. 135–150). Hershey,
PA: IGI Global. doi:10.4018/978-1-4666-4434-2.ch006

Salleh, K., Ikhsan, S. O., & Ahmad, S. N. (2011). Knowledge

management enablers and knowledge sharing process: A case study of
public sector accounting organization in Malaysia . In Al-Shammari,
M. (Ed.), Knowledge management in emerging economies: Social,
organizational and cultural implementation (pp. 199–211). Hershey,
PA: IGI Global. doi:10.4018/978-1-61692-886-5.ch012

Saunders, C. (2011). Knowledge sharing in legal practice . In

Schwartz, D., & Te’eni, D. (Eds.), Encyclopedia of knowledge
management (2nd ed.; pp. 935–945). Hershey, PA: IGI Global.
doi:10.4018/978-1-59904-931-1.ch089

Scarso, E., & Bolisani, E. (2011). Knowledge intermediation . In

Schwartz, D., & Te’eni, D. (Eds.), Encyclopedia of knowledge
management (2nd ed.; pp. 601–611). Hershey, PA: IGI Global.
doi:10.4018/978-1-59904-931-1.ch058
Scarso, E., & Bolisani, E. (2011). Managing professions for
knowledge management. International Journal of Knowledge
Management , 7(3), 61–75. doi:10.4018/jkm.2011070105

Scarso, E., & Bolisani, E. (2013). Managing professions for

knowledge management . In Jennex, M. (Ed.), Dynamic models for
knowledge-driven organizations (pp. 238–253). Hershey, PA: IGI
Global. doi:10.4018/978-1-4666-2485-6.ch014

Scarso, E., Bolisani, E., & Padova, A. (2011). The complex issue of
measuring KM performance: Lessons from the practice . In Vallejo-
Alonso, B., Rodriguez-Castellanos, A., & Arregui-Ayastuy, G. (Eds.),
Identifying, measuring, and valuing knowledge-based intangible
assets: New perspectives (pp. 208–230). Hershey, PA: IGI Global.
doi:10.4018/978-1-60960-054-9.ch010

Schumann, C., & Tittmann, C. (2010). Potentials for externalizing and

measuring of tacit knowledge within knowledge nodes in the context
of knowledge networks . In Harorimana, D. (Ed.), Cultural
implications of knowledge sharing, management and transfer:
Identifying competitive advantage (pp. 84–107). Hershey, PA: IGI
Global. doi:10.4018/978-1-60566-790-4.ch005

Schwartz, D. (2011). An Aristotelian view of knowledge for

knowledge management . In Schwartz, D., & Te’eni, D. (Eds.),
Encyclopedia of knowledge management (2nd ed.; pp. 39–48).
Hershey, PA: IGI Global. doi:10.4018/978-1-59904-931-1.ch005

Senaratne, S., & Victoria, M. F. (2014). Building a supportive culture

for sustained organisational learning in public sectors . In Al-Bastaki,
Y., & Shajera, A. (Eds.), Building a competitive public sector with
knowledge management strategy (pp. 118–134). Hershey, PA: IGI
Global. doi:10.4018/978-1-4666-4434-2.ch005
Shah, A., Singhera, Z., & Ahsan, S. (2011). Web services for
bioinformatics . In Al-Mutairi, M., & Mohammed, L. (Eds.), Cases on
ICT utilization, practice and solutions: Tools for managing day-to-day
issues (pp. 28–46). Hershey, PA: IGI Global. doi:10.4018/978-1-
60960-015-0.ch003

Shajera, A., & Al-Bastaki, Y. (2014). Organisational readiness for

knowledge management: Bahrain public sector case study . In Al-
Bastaki, Y., & Shajera, A. (Eds.), Building a competitive public sector
with knowledge management strategy (pp. 104–117). Hershey, PA:
IGI Global. doi:10.4018/978-1-4666-4434-2.ch004

Sharma, A. K., Goswami, A., & Gupta, D. (2011). An extended

relational model & SQL for fuzzy multidatabases . In Yan, L., & Ma,
Z. (Eds.), Advanced database query systems: Techniques, applications
and technologies (pp. 185–219). Hershey, PA: IGI Global.
doi:10.4018/978-1-60960-475-2.ch008

Sharma, R., Banati, H., & Bedi, P. (2012). Building socially-aware e-

learning systems through knowledge management. International
Journal of Knowledge Management , 8(3), 1–26.
doi:10.4018/jkm.2012070101

Sharma, R. S., Chandrasekar, G., & Vaitheeswaran, B. (2012). A

knowledge framework for development: Empirical investigation of 30
societies . In Jennex, M. (Ed.), Conceptual models and outcomes of
advancing knowledge management: New technologies (pp. 244–265).
Hershey, PA: IGI Global. doi:10.4018/978-1-4666-0035-5.ch013

Shaw, D. (2011). Mapping group knowledge . In Schwartz, D., &

Te’eni, D. (Eds.), Encyclopedia of knowledge management (2nd ed.;
pp. 1072–1081). Hershey, PA: IGI Global. doi:10.4018/978-1-59904-
931-1.ch102
Sheluhin, O. I., & Atayero, A. A. (2013). Principles of modeling in
information communication systems and networks . In Atayero, A., &
Sheluhin, O. (Eds.), Integrated models for information communication
systems and networks: Design and development (pp. 1–15). Hershey,
PA: IGI Global. doi:10.4018/978-1-4666-2208-1.ch001

Sheluhin, O. I., & Garmashev, A. V. (2013). Numerical methods of

multifractal analysis in information communication systems and
networks . In Atayero, A., & Sheluhin, O. (Eds.), Integrated models
for information communication systems and networks: Design and
development (pp. 16–46). Hershey, PA: IGI Global. doi:10.4018/978-
1-4666-2208-1.ch002

Siau, K., Long, Y., & Ling, M. (2010). Toward a unified model of
information systems development success. Journal of Database
Management , 21(1), 80–101. doi:10.4018/jdm.2010112304

Siau, K., Long, Y., & Ling, M. (2012). Toward a unified model of
information systems development success . In Siau, K. (Ed.), Cross-
disciplinary models and applications of database management:
Advancing approaches (pp. 80–102). Hershey, PA: IGI Global.
doi:10.4018/978-1-61350-471-0.ch004

Simard, A. J., & Jourdeuil, P. (2014). Knowledge manageability: A

new paradigm . In Al-Bastaki, Y., & Shajera, A. (Eds.), Building a
competitive public sector with knowledge management strategy (pp.
1–52). Hershey, PA: IGI Global. doi:10.4018/978-1-4666-4434-
2.ch001

Simonette, M. J., & Spina, E. (2014). Enabling IT innovation through

soft systems engineering . In Pańkowska, M. (Ed.), Frameworks of IT
prosumption for business development (pp. 64–72). Hershey, PA: IGI
Global. doi:10.4018/978-1-4666-4313-0.ch005
Sivaramakrishnan, S., Delbaere, M., Zhang, D., & Bruning, E. (2012).
Critical success factors and outcomes of market knowledge
management: A conceptual model and empirical evidence . In Jennex,
M. (Ed.), Conceptual models and outcomes of advancing knowledge
management: New technologies (pp. 165–185). Hershey, PA: IGI
Global. doi:10.4018/978-1-4666-0035-5.ch009

Small, C. T., & Sage, A. P. (2010). A complex adaptive systems-

based enterprise knowledge sharing model . In Paradice, D. (Ed.),
Emerging systems approaches in information technologies: Concepts,
theories, and applications (pp. 137–155). Hershey, PA: IGI Global.
doi:10.4018/978-1-60566-976-2.ch009

Smedlund, A. (2011). Social network structures for explicit, tacit and

potential knowledge . In Jennex, M. (Ed.), Global aspects and cultural
perspectives on knowledge management: Emerging dimensions (pp.
81–90). Hershey, PA: IGI Global. doi:10.4018/978-1-60960-555-
1.ch006

Smith, A. D. (2013). Competitive uses of information and knowledge

management tools: Case study of supplier-side management.
International Journal of Knowledge-Based Organizations , 3(1), 71–
87. doi:10.4018/ijkbo.2013010105

Smith, P., & Coakes, E. (2011). Exploiting KM in support of

innovation and change . In Eardley, A., & Uden, L. (Eds.), Innovative
knowledge management: Concepts for organizational creativity and
collaborative design (pp. 242–252). Hershey, PA: IGI Global.
doi:10.4018/978-1-60566-701-0.ch015

Smith, T. A., Mills, A. M., & Dion, P. (2010). Linking business

strategy and knowledge management capabilities for organizational
effectiveness. International Journal of Knowledge Management , 6(3),
22–43. doi:10.4018/jkm.2010070102
Smith, T. A., Mills, A. M., & Dion, P. (2012). Linking business
strategy and knowledge management capabilities for organizational
effectiveness . In Jennex, M. (Ed.), Conceptual models and outcomes
of advancing knowledge management: New technologies (pp. 186–
207). Hershey, PA: IGI Global. doi:10.4018/978-1-4666-0035-
5.ch010

Smuts, H., van der Merwe, A., & Loock, M. (2011). Key
characteristics relevant for selecting knowledge management software
tools . In Eardley, A., & Uden, L. (Eds.), Innovative knowledge
management: Concepts for organizational creativity and collaborative
design (pp. 18–39). Hershey, PA: IGI Global. doi:10.4018/978-1-
60566-701-0.ch002

Soffer, P., & Kaner, M. (2013). Complementing business process

verification by validity analysis: A theoretical and empirical
evaluation . In Siau, K. (Ed.), Innovations in database design, web
applications, and information systems management (pp. 265–288).
Hershey, PA: IGI Global. doi:10.4018/978-1-4666-2044-5.ch010

Soffer, P., Kaner, M., & Wand, Y. (2012). Assigning ontological

meaning to workflow nets . In Siau, K. (Ed.), Cross-disciplinary
models and applications of database management: Advancing
approaches (pp. 209–244). Hershey, PA: IGI Global.
doi:10.4018/978-1-61350-471-0.ch009

Sohrabi, B., Raeesi, I., & Khanlari, A. (2010). Intellectual capital

components, measurement and management: A literature survey of
concepts and measures . In López Sáez, P., Castro, G., Navas López,
J., & Delgado Verde, M. (Eds.), Intellectual capital and technological
innovation: Knowledge-based theory and practice (pp. 1–38).
Hershey, PA: IGI Global. doi:10.4018/978-1-61520-875-3.ch001

Sohrabi, B., Raeesi, I., Khanlari, A., & Forouzandeh, S. (2011). A

comprehensive model for assessing the organizational readiness of
knowledge management . In Jennex, M. (Ed.), Global aspects and
cultural perspectives on knowledge management: Emerging
dimensions (pp. 30–48). Hershey, PA: IGI Global. doi:10.4018/978-1-
60960-555-1.ch003

Sparrow, J. (2011). Knowledge management in small and medium

sized enterprises . In Schwartz, D., & Te’eni, D. (Eds.), Encyclopedia
of knowledge management (2nd ed.; pp. 671–681). Hershey, PA: IGI
Global. doi:10.4018/978-1-59904-931-1.ch065

Stam, C. D. (2011). Making sense of knowledge productivity . In

Vallejo-Alonso, B., Rodriguez-Castellanos, A., & Arregui-Ayastuy,
G. (Eds.), Identifying, measuring, and valuing knowledge-based
intangible assets: New perspectives (pp. 133–155). Hershey, PA: IGI
Global. doi:10.4018/978-1-60960-054-9.ch007

Stamkopoulos, K., Pitoura, E., Vassiliadis, P., & Zarras, A. (2012).

Accelerating web service workflow execution via intelligent
allocation of services to servers . In Siau, K. (Ed.), Cross-disciplinary
models and applications of database management: Advancing
approaches (pp. 385–416). Hershey, PA: IGI Global.
doi:10.4018/978-1-61350-471-0.ch016

Sterling, L. (2011). Applying agents within knowledge management .

In Schwartz, D., & Te’eni, D. (Eds.), Encyclopedia of knowledge
management (2nd ed.; pp. 12–19). Hershey, PA: IGI Global.
doi:10.4018/978-1-59904-931-1.ch002

Su, S., & Chiong, R. (2011). Business intelligence . In Schwartz, D.,

& Te’eni, D. (Eds.), Encyclopedia of knowledge management (2nd
ed.; pp. 72–80). Hershey, PA: IGI Global. doi:10.4018/978-1-59904-
931-1.ch008
Su, W. B., Li, X., & Chow, C. W. (2012). Exploring the extent and
impediments of knowledge sharing in Chinese business enterprise . In
Jennex, M. (Ed.), Conceptual models and outcomes of advancing
knowledge management: New technologies (pp. 266–290). Hershey,
PA: IGI Global. doi:10.4018/978-1-4666-0035-5.ch014

Subramanian, D. V., & Geetha, A. (2012). Application of

multidimensional metric model, database, and WAM for KM system
evaluation. International Journal of Knowledge Management , 8(4),
1–21. doi:10.4018/jkm.2012100101

Surendran, A., & Samuel, P. (2013). Knowledge-based code clone

approach in embedded and real-time systems . In Saeed, S., &
Alsmadi, I. (Eds.), Knowledge-based processes in software
development (pp. 49–62). Hershey, PA: IGI Global. doi:10.4018/978-
1-4666-4229-4.ch004

Takahashi, Y. (2011). The importance of balancing knowledge

protection and knowledge interchange . In Morais da Costa, G. (Ed.),
Ethical issues and social dilemmas in knowledge management:
Organizational innovation (pp. 180–198). Hershey, PA: IGI Global.
doi:10.4018/978-1-61520-873-9.ch011

Talet, A. N., Alhawari, S., & Alryalat, H. (2012). The outcome of

knowledge process for customers of Jordanian companies on the
achievement of customer knowledge retention . In Jennex, M. (Ed.),
Conceptual models and outcomes of advancing knowledge
management: New technologies (pp. 45–61). Hershey, PA: IGI
Global. doi:10.4018/978-1-4666-0035-5.ch003

Talet, A. N., Alhawari, S., Mansour, E., & Alryalat, H. (2011). The
practice of Jordanian business to attain customer knowledge
acquisition. International Journal of Knowledge Management , 7(2),
49–67. doi:10.4018/jkm.2011040103
Tanner, K. (2011). The role of emotional capital in organisational KM
. In Schwartz, D., & Te’eni, D. (Eds.), Encyclopedia of knowledge
management (2nd ed.; pp. 1396–1409). Hershey, PA: IGI Global.
doi:10.4018/978-1-59904-931-1.ch133

Tauber, D., & Schwartz, D. G. (2011). Integrating knowledge

management with the systems analysis process . In Schwartz, D., &
Te’eni, D. (Eds.), Encyclopedia of knowledge management (2nd ed.;
pp. 431–441). Hershey, PA: IGI Global. doi:10.4018/978-1-59904-
931-1.ch041

Tavana, M., Busch, T. E., & Davis, E. L. (2011). Modeling

operational robustness and resiliency with high-level Petri nets.
International Journal of Knowledge-Based Organizations , 1(2), 17–
38. doi:10.4018/ijkbo.2011040102

Taxén, L. (2010). Aligning business and knowledge strategies: A

practical approach for aligning business and knowledge strategies . In
Russ, M. (Ed.), Knowledge management strategies for business
development (pp. 277–308). Hershey, PA: IGI Global.
doi:10.4018/978-1-60566-348-7.ch013

Te’eni, D. (2011). Knowledge for communicating knowledge . In

Schwartz, D., & Te’eni, D. (Eds.), Encyclopedia of knowledge
management (2nd ed.; pp. 560–569). Hershey, PA: IGI Global.
doi:10.4018/978-1-59904-931-1.ch054

Toiviainen, H., & Kerosuo, H. (2013). Development curriculum for

knowledge-based organizations: Lessons from a learning network.
International Journal of Knowledge-Based Organizations , 3(3), 1–18.
doi:10.4018/ijkbo.2013070101

Tran, B. (2014). The human element of the knowledge worker:

Identifying, managing, and protecting the intellectual capital within
knowledge management . In Chilton, M., & Bloodgood, J. (Eds.),
Knowledge management and competitive advantage: Issues and
potential solutions (pp. 281–303). Hershey, PA: IGI Global.
doi:10.4018/978-1-4666-4679-7.ch017

Tsamoura, E., Gounaris, A., & Manolopoulos, Y. (2011). Optimal

service ordering in decentralized queries over web services.
International Journal of Knowledge-Based Organizations , 1(2), 1–16.
doi:10.4018/ijkbo.2011040101

Tsamoura, E., Gounaris, A., & Manolopoulos, Y. (2013). Optimal

service ordering in decentralized queries over web services . In Wang,
J. (Ed.), Intelligence methods and systems advancements for
knowledge-based business (pp. 43–58). Hershey, PA: IGI Global.
doi:10.4018/978-1-4666-1873-2.ch003

Tull, J. (2013). Slow knowledge: The case for savouring learning and
innovation . In Buckley, S., & Jakovljevic, M. (Eds.), Knowledge
management innovations for interdisciplinary education:
Organizational applications (pp. 274–297). Hershey, PA: IGI Global.
doi:10.4018/978-1-4666-1969-2.ch014

Turner, G., & Minonne, C. (2013). Effective knowledge management

through measurement . In Buckley, S., & Jakovljevic, M. (Eds.),
Knowledge management innovations for interdisciplinary education:
Organizational applications (pp. 145–176). Hershey, PA: IGI Global.
doi:10.4018/978-1-4666-1969-2.ch008

Uden, L., & Eardley, A. (2011). Knowledge sharing in the learning

process: Experience with problem-based learning . In Eardley, A., &
Uden, L. (Eds.), Innovative knowledge management: Concepts for
organizational creativity and collaborative design (pp. 215–229).
Hershey, PA: IGI Global. doi:10.4018/978-1-60566-701-0.ch013
Upadhyaya, S., Rao, H. R., & Padmanabhan, G. (2011). Secure
knowledge management . In Schwartz, D., & Te’eni, D. (Eds.),
Encyclopedia of knowledge management (2nd ed.; pp. 1429–1437).
Hershey, PA: IGI Global. doi:10.4018/978-1-59904-931-1.ch136

Urbancová, H., & Königová, M. (2013). The influence of the

application of business continuity management, knowledge
management, and knowledge continuity management on the
innovation in organizations . In Buckley, S., & Jakovljevic, M. (Eds.),
Knowledge management innovations for interdisciplinary education:
Organizational applications (pp. 254–273). Hershey, PA: IGI Global.
doi:10.4018/978-1-4666-1969-2.ch013

Van Canh, T., & Zyngier, S. (2014). Using ERG theory as a lens to
understand the sharing of academic tacit knowledge: Problems and
issues in developing countries – Perspectives from Vietnam . In
Chilton, M., & Bloodgood, J. (Eds.), Knowledge management and
competitive advantage: Issues and potential solutions (pp. 174–201).
Hershey, PA: IGI Global. doi:10.4018/978-1-4666-4679-7.ch010

Vat, K. H. (2011). Knowledge synthesis framework . In Schwartz, D.,

& Te’eni, D. (Eds.), Encyclopedia of knowledge management (2nd
ed.; pp. 955–966). Hershey, PA: IGI Global. doi:10.4018/978-1-
59904-931-1.ch091

Vert, S. (2012). Extensions of web browsers useful to knowledge

workers . In Jouis, C., Biskri, I., Ganascia, J., & Roux, M. (Eds.),
Next generation search engines: Advanced models for information
retrieval (pp. 239–273). Hershey, PA: IGI Global. doi:10.4018/978-1-
4666-0330-1.ch011

Wagner, L., & Van Belle, J. (2011). Web mining for strategic
competitive intelligence: South African experiences and a practical
methodology . In Al-Shammari, M. (Ed.), Knowledge management in
emerging economies: Social, organizational and cultural
implementation (pp. 1–19). Hershey, PA: IGI Global.
doi:10.4018/978-1-61692-886-5.ch001

Wautelet, Y., Schinckus, C., & Kolp, M. (2010). Towards knowledge

evolution in software engineering: An epistemological approach.
International Journal of Information Technologies and Systems
Approach , 3(1), 21–40. doi:10.4018/jitsa.2010100202

Weiß, S., Makolm, J., Ipsmiller, D., & Egger, N. (2011). DYONIPOS:
Proactive knowledge supply . In Jennex, M., & Smolnik, S. (Eds.),
Strategies for knowledge management success: Exploring
organizational efficacy (pp. 277–287). Hershey, PA: IGI Global.
doi:10.4018/978-1-60566-709-6.ch015

Welschen, J., Todorova, N., & Mills, A. M. (2012). An investigation

of the impact of intrinsic motivation on organizational knowledge
sharing. International Journal of Knowledge Management , 8(2), 23–
42. doi:10.4018/jkm.2012040102

Wickramasinghe, N. (2011). Knowledge creation . In Schwartz, D., &

Te’eni, D. (Eds.), Encyclopedia of knowledge management (2nd ed.;
pp. 527–538). Hershey, PA: IGI Global. doi:10.4018/978-1-59904-
931-1.ch051

Wickramasinghe, N. (2011). Knowledge management: The key to

delivering superior healthcare solutions . In Eardley, A., & Uden, L.
(Eds.), Innovative knowledge management: Concepts for
organizational creativity and collaborative design (pp. 190–203).
Hershey, PA: IGI Global. doi:10.4018/978-1-60566-701-0.ch011

Wijnhoven, F. (2011). Operational knowledge management . In

Schwartz, D., & Te’eni, D. (Eds.), Encyclopedia of knowledge
management (2nd ed.; pp. 1237–1249). Hershey, PA: IGI Global.
doi:10.4018/978-1-59904-931-1.ch118

Williams, R. (2011). A Knowledge Process Cycle . In Schwartz, D., &

Te’eni, D. (Eds.), Encyclopedia of knowledge management (2nd ed.;
pp. 853–866). Hershey, PA: IGI Global. doi:10.4018/978-1-59904-
931-1.ch082

Wilson, R. L., Rosen, P. A., & Al-Ahmadi, M. S. (2011). Knowledge

structure and data mining techniques . In Schwartz, D., & Te’eni, D.
(Eds.), Encyclopedia of knowledge management (2nd ed.; pp. 946–
954). Hershey, PA: IGI Global. doi:10.4018/978-1-59904-931-
1.ch090

Wimmer, H., Yoon, V., & Rada, R. (2013). Integrating knowledge

sources: An ontological approach. International Journal of Knowledge
Management , 9(1), 60–75. doi:10.4018/jkm.2013010104

Woods, S., Poteet, S. R., Kao, A., & Quach, L. (2011). Knowledge
dissemination in portals . In Schwartz, D., & Te’eni, D. (Eds.),
Encyclopedia of knowledge management (2nd ed.; pp. 539–548).
Hershey, PA: IGI Global. doi:10.4018/978-1-59904-931-1.ch052

Worden, D. (2010). Agile alignment of enterprise execution

capabilities with strategy . In Russ, M. (Ed.), Knowledge management
strategies for business development (pp. 45–62). Hershey, PA: IGI
Global. doi:10.4018/978-1-60566-348-7.ch003

Wu, D. (2010). Who are effective time managers? Bivariate

correlation analysis and hypotheses testing . In Wu, D. (Ed.),
Temporal structures in individual time management: Practices to
enhance calendar tool design (pp. 116–138). Hershey, PA: IGI Global.
doi:10.4018/978-1-60566-776-8.ch009

Wu, J., Du, H., Li, X., & Li, P. (2010). Creating and delivering a
successful knowledge management strategy . In Russ, M. (Ed.),
Knowledge management strategies for business development (pp.
261–276). Hershey, PA: IGI Global. doi:10.4018/978-1-60566-348-
7.ch012

Wu, J., Liu, N., & Xuan, Z. (2013). Simulation on knowledge transfer
processes from the perspectives of individual’s mentality and behavior
. In Yang, G. (Ed.), Multidisciplinary studies in knowledge and
systems science (pp. 233–246). Hershey, PA: IGI Global.
doi:10.4018/978-1-4666-3998-0.ch016

Wu, J., Wang, S., & Pan, D. (2013). Evaluation of technological

influence power of enterprises through the enterprise citation network
. In Yang, G. (Ed.), Multidisciplinary studies in knowledge and
systems science (pp. 34–44). Hershey, PA: IGI Global.
doi:10.4018/978-1-4666-3998-0.ch003

Wu, S. T. (2011). Innovation in new technology and knowledge

management: Comparative case studies of its evolution during a
quarter century of change . In Eardley, A., & Uden, L. (Eds.),
Innovative knowledge management: Concepts for organizational
creativity and collaborative design (pp. 77–93). Hershey, PA: IGI
Global. doi:10.4018/978-1-60566-701-0.ch005

Xiao, L., & Pei, Y. (2013). A task context aware physical distribution
knowledge service system . In Yang, G. (Ed.), Multidisciplinary
studies in knowledge and systems science (pp. 18–33). Hershey, PA:
IGI Global. doi:10.4018/978-1-4666-3998-0.ch002

Xu, D., & Wang, H. (2011). Integration of knowledge management

and e-learning . In Schwartz, D., & Te’eni, D. (Eds.), Encyclopedia of
knowledge management (2nd ed.; pp. 442–451). Hershey, PA: IGI
Global. doi:10.4018/978-1-59904-931-1.ch042
Yaniv, E., & Schwartz, D. G. (2011). Organizational attention . In
Schwartz, D., & Te’eni, D. (Eds.), Encyclopedia of knowledge
management (2nd ed.; pp. 1270–1279). Hershey, PA: IGI Global.
doi:10.4018/978-1-59904-931-1.ch121

Yeung, C. L., Cheung, C. F., Wang, W. M., & Tsui, E. (2013). A

study of organizational narrative simulation for decision support . In
Yang, G. (Ed.), Multidisciplinary studies in knowledge and systems
science (pp. 179–192). Hershey, PA: IGI Global. doi:10.4018/978-1-
4666-3998-0.ch013

Yigitcanlar, T. (2013). Moving towards a knowledge city? Brisbane’s

experience in knowledge-based urban development . In Wang, J.
(Ed.), Intelligence methods and systems advancements for
knowledge-based business (pp. 114–131). Hershey, PA: IGI Global.
doi:10.4018/978-1-4666-1873-2.ch007

Yigitcanlar, T., & Martinez-Fernandez, C. (2010). Making space and

place for knowledge production: Socio-spatial development of
knowledge community precincts . In Metaxiotis, K., Carrillo, F., &
Yigitcanlar, T. (Eds.), Knowledge-based development for cities and
societies: Integrated multi-level approaches (pp. 99–117). Hershey,
PA: IGI Global. doi:10.4018/978-1-61520-721-3.ch006

Yıldırım, A. A., Özdoğan, C., & Watson, D. (2014). Parallel data

reduction techniques for big datasets . In Hu, W., & Kaabouch, N.
(Eds.), Big data management, technologies, and applications (pp. 72–
93). Hershey, PA: IGI Global. doi:10.4018/978-1-4666-4699-5.ch004

Yoon, K. S. (2012). Measuring the influence of expertise and

epistemic engagement to the practice of knowledge management.
International Journal of Knowledge Management , 8(1), 40–70.
doi:10.4018/jkm.2012010103
Yusof, Z. M., & Ismail, M. B. (2011). Factors affecting knowledge
sharing practice in Malaysia: A preliminary overview . In Al-
Shammari, M. (Ed.), Knowledge management in emerging
economies: Social, organizational and cultural implementation (pp.
157–170). Hershey, PA: IGI Global. doi:10.4018/978-1-61692-886-
5.ch009

Zapata-Cantú, L., Ramírez, J., & Pineda, J. L. (2011). HRM

adaptation to knowledge management initiatives: Three Mexican
cases . In Al-Shammari, M. (Ed.), Knowledge management in
emerging economies: Social, organizational and cultural
implementation (pp. 273–293). Hershey, PA: IGI Global.
doi:10.4018/978-1-61692-886-5.ch017

Zarri, G. P. (2011). Knowledge representation . In Schwartz, D., &

Te’eni, D. (Eds.), Encyclopedia of knowledge management (2nd ed.;
pp. 878–892). Hershey, PA: IGI Global. doi:10.4018/978-1-59904-
931-1.ch084

Zarri, G. P. (2011). RDF and OWL for knowledge management . In

Schwartz, D., & Te’eni, D. (Eds.), Encyclopedia of knowledge
management (2nd ed.; pp. 1355–1373). Hershey, PA: IGI Global.
doi:10.4018/978-1-59904-931-1.ch130

Zhang, Y., Wang, Y., Colucci, W., & Wang, Z. (2013). The paradigm
shift in organizational research . In Wang, J. (Ed.), Intelligence
methods and systems advancements for knowledge-based business
(pp. 60–74). Hershey, PA: IGI Global. doi:10.4018/978-1-4666-1873-
2.ch004

Zhang, Z. J. (2011). Managing customer knowledge with social

software . In Schwartz, D., & Te’eni, D. (Eds.), Encyclopedia of
knowledge management (2nd ed.; pp. 1046–1053). Hershey, PA: IGI
Global. doi:10.4018/978-1-59904-931-1.ch099
Zyngier, S. (2011). Governance of knowledge management . In
Schwartz, D., & Te’eni, D. (Eds.), Encyclopedia of knowledge
management (2nd ed.; pp. 354–365). Hershey, PA: IGI Global.
doi:10.4018/978-1-59904-931-1.ch034

Zyngier, S. (2011). Knowledge management: Realizing value through

governance. International Journal of Knowledge Management , 7(1),
35–54. doi:10.4018/jkm.2011010103

Zyngier, S. (2013). Knowledge management: Realizing value through

governance . In Jennex, M. (Ed.), Dynamic models for knowledge-
driven organizations (pp. 36–55). Hershey, PA: IGI Global.
doi:10.4018/978-1-4666-2485-6.ch003
Compilation of References

A Survey on Feature Selection Techniques and Classification

Algorithms for Efficient Text Classification. (2016). International
Journal of Science and Research.

Abdalla, M., Bellare, M., Catalano, D., Kiltz, E., Kohno, T., Lange,
T., & Shi, H. (2008). Searchable encryption revisited: Consistency
properties, relation to anonymous ibe, and extensions . Journal of
Cryptology , 21(3), 350–391. doi:10.1007/s00145-007-9006-6

Alavi, A. H., & Gandomi, A. H. (2011). A robust data mining

approach for formulation of geotechnical engineering systems.
Engineering Computations , 28(3), 242–274.
doi:10.1108/02644401111118132

Ali AlShaari, M. (2014). Text Documents Classification Using Word

Intersections. International Journal of Engineering and Technology ,
6(2), 119–122. doi:10.7763/IJET.2014.V6.678

Alicherry, M., & Lakshman, T. V. (2012). Network aware resource

allocation in distributed clouds . Proceedings - IEEE INFOCOM , 12,
963–971.

Alliance, C. S. (2009). Security Guidance for Critical Areas of Focus

in Cloud Computing. Retrieved from
http://www.cloudsecurityalliance.org

Armbrust, Fox, Griffith, Joseph, & Katz, Konwinski, … Zaharia.

(2010). A view of cloud computing. Communications of the ACM ,
53(4), 50–58.

Arvor, D., Jonathan, M., Meirelles, M. S. P., Dubreuil, V., & Durieux,
L. (2011). Classification of MODIS EVI time series for crop mapping
in the state of MatoGrosso. Brazil International Journal of Remote
Sensing , 32(22), 7847–7871. doi:10.1080/01431161.2010.531783

Austin. (2012). eDiscovery Trends: CGOCs Information Lifecycle

Governance Leader Reference Guide. Retrieved from
http://www.ediscoverydaily.com

Bakshi K. (2012). Considerations for big data: Architecture and

approach. Aerospace Conference, 1–7. 10.1109/AERO.2012.6187357

Ballard L. Kamara S. Monrose F. (2005). Achieving efficient

conjunctive keyword searches over encrypted data. Proc. of ICICS.
10.1007/11602897_35

Barham P. Donnelly A. Isaacs R. Mortier R. (2004). Using magpie for

request extraction and workload modelling. Proceedings of the 6th
conference on Symposium on Opearting Systems Design &
Implementation (OSDI’04), 18–18.

Basori, Afif, Almazyad, Abujabal, Rehman, & Alkawaz. (2015). Fast

Markerless Tracking for Augmented Reality in Planar Environment.
3D Research, 6(4), 1-11.

Bellare M. Boldyreva A. Neill A. O. (2007). Deterministic and

efficiently searchable encryption. Proc. of CRYPTO.

Berry, M. J., & Linoff, G. S. (1997). Data mining techniques: For

marketing, sales, and customer support . John Wiley & Sons, Inc.

Berzal, F., Cubero, J. C., Cuenca, F., & Martín-Bautista, M. J. (2003).

On the quest for easy-to-understand splitting rules. Data &
Knowledge Engineering , 44(1), 31–48. doi:10.1016/S0169-
023X(02)00062-9

Beyer & Laney. (2012). The importance of ‘big data’: A definition .

Gartner.

Big Data, Big Impact: New Possibilities for International

Development. (n.d.). World Economic Forum.

Bijalwan, V., Kumar, V., Kumari, P., & Pascual, J. (2014). KNN
based machine learning approach for text and document mining.
International Journal of Database Theory and Application , 7(1), 61–
70. doi:10.14257/ijdta.2014.7.1.06

Bill Franks. (2012). Taming the big data tidal wave . Wiley.

Boldyreva, A., Chenette, N., Lee, Y., & Oneill, A. (2009). Order-
preserving symmetric encryption . In Advances in Cryptology-
EUROCRYPT (pp. 224–241). Springer.

Boneh, Kushilevitz, Ostrovsky, & W. E. S. III. (2007). Public key

encryption that allows pir queries. Proc. of CRYPTO.

Boneh D. Crescenzo G. D. Ostrovsky R. Persiano G. (2004). Public

key encryption with keyword search. Proc. of EUROCRYPT.

Boneh D. Waters B. (2007). Conjunctive, subset, and range queries on

encrypted data. Proc. of TCC, 535–554.

Bowyer K. W. Hall L. O. Moore T. Chawla N. Kegelmeyer W. P.

(2000), A parallel decision tree builder for mining very large
visualization datasets. IEEE International Conference on Systems
Man, and Cybernetics, 3, 1888-1893. 10.1109/ICSMC.2000.886388
Brinkman, R. (2007). Searching in encrypted data. PhD thesis.

Brinkman. (2007). Searching in encrypted data. PhD thesis.

Cantrill B. M. Shapiro M. W. Leventhal A. H. (2004). Dynamic

instrumentation of production systems. USENIX Annual Technical
Conference, 15–28.

Cao, N., Wang, C., Li, M., Ren, K., & Lou, W. (2014). Privacy-
preserving multikeyword ranked search over encrypted cloud data .
IEEE Transactions on Parallel and Distributed Systems , 25(1), 222–
233. doi:10.1109/TPDS.2013.45

Cao, Wang, & Li, Ren, & Lou. (2014). Privacy-preserving

multikeyword ranked search over encrypted cloud data. IEEE
Transactions on Parallel and Distributed Systems , 25(1), 222–233.

Cash, J., & Jutla, K., Ro3u, & Steiner. (2013). Highly-scalable
searchable symmetric encryption with support for Boolean queries.
Proc. CRYPTO, 353-373.

Chang Y.-C. Mitzenmacher M. (2005). Privacy preserving keyword

searches on remote encrypted data. Proc. of ACNS.
10.1007/11496137_30

Chen M. Y. Accardi A. Kiciman E. Lloyd J. Patterson D. Fox A.

Brewer E. (2004). Path-based faliure and evolution management. In
Proceedings of the 1st conference on Symposium on Networked
Systems Design and Implementation. USENIX Association.

Chen M. Kiciman E. Fratkin E. Fox A. Brewer E. (n.d.). Pinpoint:

problem determination in large, dynamic internet services.
Proceedings of the International Conference on Dependable Systems
and Networks (DSN’02), 595–604. 10.1109/DSN.2002.1029005
Chen, X., Zheng, Z., Liu, X., Huang, Z., & Sun, H. (2011).
Personalized QoS-aware Web service recommendation and
visualization. IEEE Transactions on Services Computing .

Cloudera's 100% Open Source Distribution of Hadoop. (n.d.).

Retrieved from http://www.cloudera.com/content/clou
dera/en/products/cdh.html

Coyle, D. J., Jr., Chang, A., Malkemus, T. R., & Wilson, W. G.

(1997). U.S. Patent No. 5,630,124. Washington, DC: U.S. Patent and
Trademark Office.

Curtmola R. Garay J. A. Kamara S. Ostrovsky R. (2006). Searchable

symmetric encryption: improved definitions and efficient
constructions. Proc. of ACM CCS. 10.1145/1180405.1180417

David, R. (2013). Getting started with business analytics – insightful

decision making . Talor & Francis Group.

Dayan, P. (n.d.). Unsupervised Learning. Appeared . In Wilson, R. A.,

& Keil, F. (Eds.), The MIT Encyclopedia Of The Cognitive Sciences .
MIT.

De Capitani di Vimercati, S., Foresti, S., & Samarati, P. (2012).

Managing and accessing data in the cloud: Privacy risks and
approaches. CRiSIS, 1 –9.

Dean, J., & Ghemawat, S. (2008). MapReduce: Simplified data

processing on large clusters . Communications of the ACM , 51(1),
107–113. doi:10.1145/1327452.1327492

DeCandia , Hastorun, Jampani, Kakulapati, Lakshman, Pilchin, …

Vogels. (2007). Dynamo: amazon’s highly available key-value store.
Proc. 21st ACM Symposium on Operating Systems Principles
(SOSP’07), 205–220.
(SOSP’07), 205–220.

Demirkan, H., & Delen, D. (2013). Leveraging the capabilities of

service-oriented decision support systems: Putting analytics and big
data in cloud . Decision Support Systems , 55(1), 412–421.
doi:10.1016/j.dss.2012.05.048

Dey A. (2016). A Systematic Review of Usability Studies in

Augmented Reality between 2005 and 2014. Mixed and Augmented
Reality (ISMAR-Adjunct), 2016 IEEE International Symposium on.

Dinh, Lee, Niyato, & Wang. (2013). A survey of mobile cloud

computing: Architecture, applications, and approaches. Wireless
Commun. Mobile Comput., 13(18).

Dong, X. L., & Srivastava, D. (2013). Big data integration. In Data

Engineering (ICDE), 2013 IEEE 29th International Conference on,
1245–1248. 10.1109/ICDE.2013.6544914

Dong, J., Xiao, X., Kou, W., Qin, Y., Zhang, G., Li, L., & Moore, B.
III. (2015). Tracking the dynamics of paddy rice planting area in
1986-2010 through time series Landsat images and phenology-based
algorithms. Remote Sensing of Environment , 160, 99–113.
doi:10.1016/j.rse.2015.01.004

Dubey, V. (2016). Hybrid classification model of correlation-based

feature selection and support vector machine . IEEE.
doi:10.1109/ICCTAC.2016.7567338

Elnikety, S., Pedone, F., & Zwaenepoel, W. (2005, October).

Database replication using generalized snapshot isolation. In Reliable
Distributed Systems, 2005. SRDS 2005. 24th IEEE Symposium on (pp.
73-84). IEEE. 10.1109/RELDIS.2005.14

Elnikety, S., Dropsho, S., & Pedone, F. (2006, April). Tashkent:

Uniting durability with transaction ordering for high-performance
scalable database replication. Operating Systems Review , 40(4), 117–
130. doi:10.1145/1218063.1217947

Eswaran, K. P., Gray, J. N., Lorie, R. A., & Traiger, I. L. (1976). The
notions of consistency and predicate locks in a database system.
Communications of the ACM , 19(11), 624–633.
doi:10.1145/360363.360369

Gayathri, K. (2013). Data pre-processing with the KNN for

classification using the SVM . IEEE.

Gentry, C. (2009). A fully homomorphic encryption scheme (Ph.D.

dissertation). Stanford University.

Global data center traffic – Cisco Forecast Overview. (n.d.). Retrieved

from http://www.cisco.com/en/US/solutions/
collateral/ns341/ns525/ns537/ns705/ns
1175/Cloud_Index_White_Paper.html

Goh, E.-J. (2003). Secure indexes. Retrieved from

http://eprint.iacr.org/2003/216

Goldreich, O., & Ostrovsky, R. (1996). Software protection and

simulation on oblivious rams . Journal of the Association for
Computing Machinery , 43(3), 431–473. doi:10.1145/233551.233553

Golle P. Staddon J. Waters B. (2004). Secure conjunctive keyword

search over encrypted data. Proc. of ACNS, 31–45. 10.1007/978-3-
540-24852-1_3

Graham, S. L., Kessler, P. B., & Mckusick, M. K. (1982). Gprof: A

call graph execution profiler . ACM SIGPLAN Notices , 17(6), 120–
126. doi:10.1145/872726.806987
Guang-chao, W. (2008). Support Vector Machine Classifier Based on
Fuzzy Partition and Neighborhood Pairs. Jisuanji Yingyong .

Gumus, F. (2014). Online Naive Bayes classification for network

intrusion detection . IEEE. doi:10.1109/ASONAM.2014.6921657

Gunaratna, Kodikara, & Premaratne. (2008). ANN based currency

recognition system using compressed gray scale and application for
Sri Lankan currency notes-SLCRec. Proceedings of World Academy
of Science, Engineering and Technology, 35, 235-240.

Guo, Zhao, & Cai. (2010). A reliable method for paper currency
recognition based on LBP. In Network Infrastructure and Digital
Content, 2010 2nd IEEE International Conference on. IEEE.

Gupta A. Agarwal D. Tan D. Kulesza J. Pathak R. Stefani S.

Srinivasan V. (2015, May). Amazon Redshift and the case for simpler
data warehouses. In Proceedings of the 2015 ACM SIGMOD
International Conference on Management of Data (pp. 1917-1923).
ACM. 10.1145/2723372.2742795

Han S. Dang Y. Ge S. Zhang D. Xie T. (2012). Performance

debugging in the large via mining millions of stack traces. Proc. 34th
Int’l Conf. on Software Engineering (ICSE’12), 145–155.
10.1109/ICSE.2012.6227198

Hassanpour, Yaseri, & Ardeshiri. (2007). Feature extraction for paper

currency recognition. In Signal Processing and Its Applications, 2007.
ISSPA 2007. 9th International Symposium on. IEEE.

Hitzler, P., & Janowicz, K. (2013). Linked Data, Big Data, and the 4th
Paradigm. Semantic Web , 4(3), 233–235.

Howe, D., Costanzo, M., Fey, P., Gojobori, T., Hannick, L., Hide, W.,
& Rhee, S. Y. (2008). Big data: The future of biocuration . Nature ,
455(7209), 47–50. doi:10.1038/455047a

Hwan, Gelogo, & Park. (2012). Next Generation Cloud Computing

Issues and Solutions. International Journal of Control and
Automation, 5.

Hwang, Y., & Lee, P. (2007). Public key encryption with conjunctive
keyword search and its extension to a multi-user system . Pairing.
doi:10.1007/978-3-540-73489-5_2

IBM. (2013). What is big data? ł bringing big data to the enterprise.
Retrieved from http://www-01.ibm.com/software/data/bigdata

Jason, J. K. P., Park, Y., & Mahlke, S. (2013). Efficient execution of

augmented reality applications on mobile programmable accelerators.
Field-Programmable Technology (FPT) 2013 International
Conference on, 176-183.

Jung T. Mao X. Li X. Tang S.-J. Gong W. Zhang L. (2013). Privacy

preserving data aggregation without secure channel: multivariate
polynomial evaluation. Proceedings of INFOCOM, 2634–2642.
10.1109/INFCOM.2013.6567071

Kamara & Lauter. (2010). Cryptographic cloud storage. In RLCPS.

Springer.

Kamara, S., & Lauter, K. (2010). Cryptographic cloud storage. In

RLCPS. Springer. doi:10.1007/978-3-642-14992-4_13

Kang Y. Zheng Z. Lyu M. (2012). A latency-aware co-deployment

mechanism for cloud-based services. Proceedings of the IEEE 5th
International Conference on Cloud Computing (CLOUD’12), 630–
637. 10.1109/CLOUD.2012.90
Katz, J., Sahai, A., & Waters, B. (2008). Predicate encryption
supporting disjunctions, polynomial equations, and inner products .
Proc. of EUROCRYPT. doi:10.1007/978-3-540-78967-3_9

Khan, Wang, Kulsoom, & Ullah. (2013). Searching Encrypted Data

on Cloud. International Journal of Computer Science Issues, 10(6).

Kristensen, B. B., May, D., & Nowack, P. (2011). Collaboration and

Modeling in Ambient Systems: Vision Concepts and Experiments.
System Sciences (HICSS) 2011 44th Hawaii International Conference
on, 1-6.

Kulesza, A. (2012). Determinantal Point Processes for Machine

Learning. Foundations And Trends® In . Machine Learning , 5(2-3),
123–286. doi:10.1561/2200000044

Kumar, S. (2004, January 01). MEGA3: Integrated software for

Molecular Evolutionary Genetics Analysis and sequence alignment.
Briefings in Bioinformatics , 5(2), 150–163. doi:10.1093/bib/5.2.150

Kumar, V., Grama, A., Gupta, A., & Karypis, G. (1994). Introduction
to parallel computing (Vol. 110). Redwood City:
Benjamin/Cummings.

Kumar, V., Grama, A., Gupta, A., & Karypis, G. (1994). Introduction
to parallel computing: design and analysis of algorithms (Vol. 400).
Redwood City, CA: Benjamin/Cummings.

Kuttiyapillai, D., & Ramachandran, R. 2014. Design and analysis of

feature classification model using information extraction in tomato
growing environment. International Information Institute (Tokyo).
Information, 17(8), 3947-3959.

Kuttiyapillai, D., & Rajeswari, R. (2015). A method for extracting

task-oriented information from biological text sources. International
Journal of Data Mining and Bioinformatics , 12(4), 387–399.
doi:10.1504/IJDMB.2015.070072

Labrinidis, A., & Jagadish, H. V. (2012). Challenges and

opportunities with big data . Proceedings of the VLDB Endowment
International Conference on Very Large Data Bases , 5(12), 2032–
2033. doi:10.14778/2367502.2367572

Laskov, P., & Lippmann, R. (2010). Machine learning in adversarial

environments. Machine Learning , 81(2), 115–119.
doi:10.1007/s10994-010-5207-6

Lewko A. Okamoto T. Sahai A. Takashima K. Waters B. (2010).

Fully secure functional encryption: Attribute-based encryption and
(hierarchical) inner product encryption. Proc. of EUROCRYPT.
10.1007/978-3-642-13190-5_4

Li, H., Dai, Y., Tian, L., & Yang, H. (2009). Identity-based
authentication for cloud computing. In Cloud Computing. Berlin,
Germany: Springer-Verlag. doi:10.1007/978-3-642-10665-1_14

Li, J., Wang, Q., Wang, C., Cao, N., Ren, K., & Lou, W. (2010).
Fuzzy keyword search over encrypted data in cloud computing. Proc.
of IEEE INFOCOM’10 Mini-Conference.
doi:10.1109/INFCOM.2010.5462196

Liang, Cai, Huang, Shen, & Peng. (2012). An SMDPbased service

model for interdomain resource allocation in mobile cloud networks.
IEEE Trans. Veh. Technol., 61(5).

Liang, H., Cai, L. X., Huang, D., Shen, X., & Peng, D. (2012). An
smdpbased service model for interdomain resource allocation in
mobile cloud networks . IEEE Transactions on Vehicular Technology
, 61(5), 2222–2232. doi:10.1109/TVT.2012.2194748

Li, H., Liu, D., Dai, Y., Luan, T. H., & Shen, X. (2014). Enabling
efficient multikeyword ranked search over encrypted cloud data
through blind storage . IEEE Transactions on Emerging Topics in
Computing . doi:doi:10.1109/TETC.2014.2371239

Lim F. Singh N. Yajnik S. (2008). A log mining approach to failure

analysis of enterprise telephony systems. Proceedings of the IEEE
International Conference on Dependable Systems and Networks
(DSN’08), 398–403.

Li, R., Xu, Z., Kang, W., Yow, K. C., & Xu, C.-Z. (2014). Efficient
multikeyword ranked query over encrypted data in cloud computing .
Future Generation Computer Systems , 30, 179–190.
doi:10.1016/j.future.2013.06.029

Liu, Li, Huang, & Wen. (2012). Shapley value based impression
propagation for reputation management in web service composition.
Pro. IEEE 19th Int’l Conf’ on Web Services (ICWS’12), 58–65.

Lo, W., Yin, J., Deng, S., Li, Y., & Wu, Z. (2012). Collaborative web
service qos prediction with location-based regularization. Pro. IEEE
19th Int’l Conf’ on Web Services (ICWS’12), 464–471.
10.1109/ICWS.2012.49

Lobell, D. B., Thau, D., Seifert, C., Engle, E., & Little, B. (2015). A
scalable satellite-based crop yield mapper. Remote Sensing of
Environment , 164, 324–333. doi:10.1016/j.rse.2015.04.021

Lyu, M. R. (1995). Software Fault Tolerance. In Trends in Software .

Wiley.

MacWilliams, Reicher, Klinker, & Bruegge. (2014). Design Patterns

for Augmented Reality Systems. CEUR Workshop Proceedings , ▪▪▪,
19.

Madsen, J. B., & Stenholt, R. (2014). How wrong can you be:
Perception of static orientation errors in mixed reality. 3D User
Interfaces (3DUI) 2014 IEEE Symposium on, 83-90.
10.1109/3DUI.2014.6798847

Mahmoud & Shen. (2012). A cloud-based scheme for protecting

source-location privacy against hotspot-locating attack in wireless
sensor networks. IEEE Trans. Parallel Distrib. Syst., 23(10).

Mahmoud, M. M., & Shen, X. (2012). A cloud-based scheme for

protecting source-location privacy against hotspot-locating attack in
wireless sensor networks . IEEE Transactions on Parallel and
Distributed Systems , 23(10), 1805–1818.
doi:10.1109/TPDS.2011.302

Manyika, Chui, Brown, Bughin, Dobbs, Roxburgh, & Byers. (2011).

Big data: The next frontier for innovation, competition, and
productivity. McKinsey Global Institute.

Mark, A. (2012). Beyer and Douglas Laney. “The Importance of ‘Big

Data’: A Definition . Gartner.

Mell & Grance. (2011). The nist definition of cloud computing (draft).
NIST Special Publication, 800, 145.

Metropolis, N. (1986). Massively parallel processing. Journal of

Scientific Computing , 1(2), 115–116. doi:10.1007/BF01061388

Mi H. Wang H. Cai H. Zhou Y. Lyu M. R. Chen Z. (2012). Ptracer:

Path-based performance profiling in cloud computing systems. In
Proceedings of the 36th IEEE Annual Computer Software and
Applications Conference (COMPSAC’12). IEEE.
Applications Conference (COMPSAC’12). IEEE.

Minka, T. P., & Picard, R. W. (1997). Interactive learning using a

“society of models”. Pattern Recognition , 30(4), 565–581.
doi:10.1016/S0031-3203(96)00113-6

Mi, Wang, & Zhou, Lyu, & Cai. (2013). Towards fine-grained,
unsupervised, scalable performance diagnosis for production cloud
computing systems. IEEE Transactions on Parallel and Distributed
Systems .

Naveed, Prabhakaran, & Gunter. (2014). Dynamic searchable

encryption via blind storage. Proceedings - IEEE Symposium on
Security and Privacy , ▪▪▪, 639–654.

Nguyen, B. V., Pham, D., Ngo, T. D., Le, D. D., & Duong, D. A.
(2014, December). Integrating spatial information into inverted index
for large-scale image retrieval. In Multimedia (ISM), 2014 IEEE
International Symposium on (pp. 102-105). IEEE. doi:10.1007/978-3-
319-12024-9_19

OECD. (2015). Data-driven innovation: big data for growth and well-
being . Paris, France: OECD Publishing.

Oracle Big Data strategy guide. (n.d.). Retrieved from

http://www.oracle.com/us/technologies bigdatabigdata-strategy-
guide- 1536569.pdf

Paillier, P. (1999). Public key cryptosystems based on composite

degree residuosity classes. Eurocrypt, 223–238. doi:10.1007/3-540-
48910-X_16

Pang, Shen, & Krishnan. (n.d.). Privacy-preserving similarity-based

text retrieval. ACM Transactions on Internet Technology, 10(1), 4.
Polo, J. L., Berzal, F., & Cubero, J. C. (2007). Taking class
importance into account. Lecture Notes in Computer Science , 4413.

Provost & Fawcett. (2013). Data science for business. O’Reilly.

Quinlan. (1996). Improved use of continuous attributes in C4.5. arXiv

preprint cs/9603103

Quinlan, J. R. (1986). Induction of decision trees . Machine Learning ,

1(1), 81–106. doi:10.1007/BF00116251

Quinlan, J. R. (1990). Decision trees and decision-making . IEEE

Transactions on Systems, Man, and Cybernetics , 20(2), 339–346.
doi:10.1109/21.52545

Quinlan, J. R. (1993). C4.5: programs for machine learning . Morgan

Kaufmann.

Ramesh, B. (2015). Big data architecture . In Big Data (pp. 29–59).

Springer. doi:10.1007/978-81-322-2494-5_2

Reddy. (2013). Techniques for Efficient Keyword Search in Cloud

Computing. International Journal of Computer Science and
Information Technologies, 4(1).

Reges, S., & Stepp, M. (2014). Building Java Programs . Pearson.

Ren, K., Wang, C., & Wang, Q. (2012). Security Challenges for the
Public Cloud . IEEE Internet Computing , 16(1), 69–73.
doi:10.1109/MIC.2012.14

Reynolds P. Killian C. Wiener J. L. Mogul J. C. Shah M. A. Vahdat

A. (2006). Pip: detecting the unexpected in distributed systems.
Proceedings of the 3rd conference on Networked Systems Design &
Implementation (NSDI’06), 9–9.
Ristoski, P., Mencía, E. L., & Paulheim, H. (2014, May). A hybrid
multi-strategy recommender system using linked open data. In
Semantic Web Evaluation Challenge (pp. 150-156). Springer
International Publishing. doi:10.1007/978-3-319-12024-9_19

Ristoski, P., Mencía, E. L., & Paulheim, H. (2014, May). A hybrid

multi-strategy recommender system using linked open data . In
Semantic Web Evaluation Challenge (pp. 150–156). Springer
International Publishing.

Rui, Y., Huang, T. S., Ortega, M., & Mehrota, S. (1998). Relevance
feedback: A power tool for interactive content-based image retrieval.
IEEE Transaction on Circuits System and Video Technnology , 8(5),
644–655. doi:10.1109/76.718510

Samanthula, B. K., Elmehdwi, Y., & Jiang, W. (2014). k-nearest

neighbor classification over semantically secure encrypted relational
data. eprint arXiv:1403.5001

Sankaranarayanan, S., & Perumal, T. P. (2014). Diabetic Prognosis

through Data Mining Methods and Techniques,” Intelligent
Computing Applications (ICICA), 2014 International Conference on,
162-166.

Schatz, M. C. (2009). CloudBurst: Highly sensitive read mapping

with MapReduce. Bioinformatics (Oxford, England) , 25(11), 1363–
1369. doi:10.1093/bioinformatics/btp236

Senthil, G. (2004). U.S. Patent No. 6,721,869. Washington, DC: U.S.

Patent and Trademark Office.

Shafer J. Agrawal R. Mehta M. (1996). PRINT: A scalable parallel

classifier for data mining. Proc. Int.Conf. Very Large Data Bases.
Shao L. Zhang J. Wei Y. Zhao J. Xie B. Mei H. (2007). Personalized
QoS prediction for Web services via collaborative filtering. Proc. 5th
Int’l Conf. Web Services (ICWS’07), 439–446.
10.1109/ICWS.2007.140

Shen, Liang, Shen, Lin, & Luo. (2014). Exploiting geodistributed

clouds for a e-health monitoring system with minimum service delay
and privacy preservation. IEEE J. Biomed. Health Inform., 18(2).

Shen E. Shi E. Waters B. (2009). Predicate privacy in encryption

systems. Proc. of TCC.

Shen, Q., Liang, X., Shen, X., Lin, X., & Luo, H. (2014). Exploiting
geodistributed clouds for e-health monitoring system with minimum
service delay and privacy preservation . IEEE Journal of Biomedical
and Health Informatics , 18(2), 430–439.
doi:10.1109/JBHI.2013.2292829

Sigelman, Barroso, Burrows, Stephenson, Plakal, Beaver, …

Shanbhag. (2010). Dapper, a large-scale distributed systems tracing
infrastructure. Google, Inc.

Singhal, A. (2001). Modern information retrieval: A brief overview .

A Quarterly Bulletin of the Computer Society of the IEEE Technical
Committee on Data Engineering , 24(4), 35–43.

Song, D. X., Wagner, D., & Perrig, A. (2000). Practical techniques for
searches on encrypted data. In Proceedings of S&P. IEEE.

Song, D. X., Wagner, D., & Perrig, A. (2000). Practical techniques for
searches on encrypted data. Proceedings of S&P, 44–55.

Song D. Wagner D. Perrig A. (2000). Practical techniques for

searches on encrypted data. Proc. of S&P.
Stefanov, Papamanthou, & Shi. (2014). Practical dynamic searchable
encryption with small leakage. Proc. NDSS.
doi:10.1109/TPDS.2013.282

Steiner M. Gaglianello B. G. Gurbani V. K. Hilt V. Roome W. D.

Scharf M. Voith T. (2012). Network-aware service placement in a
distributed cloud environment. Proc. ACM SIGCOMM’12, 73–74.
10.1145/2342356.2342366

Sun, W. (2013). Privacy-preserving multikeyword text search in the

cloud supporting similarity-based ranking. Proc. 8th ACM SIGSAC
Symp.Inf., Comput. Commun. Secur., 71-82.

Sun, Wang, Cao, Li, Lou, Hou, & Li. (2013). Verifiable privacy-
preserving multikeyword text search in the cloud supporting
similarity-based ranking. IEEE Transactions on Parallel and
Distributed Systems. DOI: 10.1109/TPDS.2013.282

Talwar, A., & Kumar, Y. (2013). Article. International Journal of

Engineering and Computer Science, 2(12).

Tang, M., Jiang, Y., Liu, J., & Liu, X. F. (2012). Location-aware
collaborative filtering for qos-based service recommendation. Pro.
IEEE 19th Int’l Conf’ on Web Services (ICWS’12), 202–209.

Taylor, R. C. (2010). An overview of the Hadoop/MapReduce/HBase

framework and its current applications in bioinformatics . BMC
Bioinformatics , 11.

Teather & Stuerzlinger. (2016). SIVARG: Spatial Interaction in

Virtual/Augmented Reality and Games. Proceedings of the 2016.
doi:10.1109/ISMAR-Adjunct.2016.0036

The Economist. (2010). A special report on managing information:

Data, data everywhere. The Economist.

Thereska, G., Salmon, B., Strunk, J., Wachs, M., Abd-El-Malek, M.,
Lopez, J., & Ganger, G. R. (2006). Stardust: tracking activity in a
distributed storage system. ACM SIGMETRICS Performance
Evaluation Review, 34(1), 3–14. doi:10.1145/1140277.1140280

Thereska H. Ganger G. R. (2008). Ironmodel: robust performance

models in the wild. Proceedings of the International Conference on
Measurement and Modeling of Computer Systems (SIGMETRICS
’08), 253–264. 10.1145/1375457.1375486

Tukey, J. W. (1977). Exploratory data analysis. Academic Press.

Valsala, S., Ann George, J., & Parvathy, P. (2011). A Study of

Clustering and Classification Algorithms Used in Datamining.
International Journal of Computer Science and Network Security,
11(10).

Vaquero, L. M., Rodero-Merino, L., Caceres, J., & Lindner, M.

(2009). A break in the clouds: Towards a cloud definition . ACM
SIGCOMM Comput. Commun. Rev. , 39(1), 50–55.
doi:10.1145/1496091.1496100

Veltkamp, C. R., Burkhardt, H., & Kriegel, H. P. (2001). State-of-the-

Art in content-based image and video retrieval . Norwell, MA:
Kluwer. doi:10.1007/978-94-015-9664-0

Venkatesan, Arunkumar, Thangavelu, & Prabhavathy. (2013). An

Improved Bayesian Classification Data mining Method for Early
Warning Landslide Susceptibility Model Using GIS. Proceedings of
Seventh International Conference on Bio-Inspired Computing:
Theories and Applications (BIC-TA 2012). Springer.
Venkatesan, , Rajawat, Arunkumar, Anbarasi, & Malarvizhi. (2014).
GIS Based Data Mining Classification Approaches for Landslide
Susceptibility Analysis . International Journal of Applied
Environmental Sciences , 9(5), 2345–2357.

Vermote, E. F., Tanre, D., Deuze, J. L., Herman, M., & Morcrette, J.
(1997). Second simulation of the satellite signal in the solar spectrum,
6S: An overview. IEEE Transactions on Geoscience and Remote
Sensing , 35(3), 675–686. doi:10.1109/36.581987

Wang C. Cao N. Li J. Ren K. Lou W. (2010). Secure ranked keyword

search over encrypted cloud data. In Proceedings of ICDCS. IEEE.
10.1109/ICDCS.2010.34

Wang, L. (2012). Improved KNN classification algorithms research in

text categorization . IEEE. doi:10.1109/CECNet.2012.6201850

Wang, S. (2013). Support Vector Machines Classification for High-

Dimensional Dataset . IEEE.

Wang, Yu, Lou, & Hou. (2014). Privacy-preserving multikeyword

fuzzy search over encrypted data in the cloud. Proceedings - IEEE
INFOCOM .

White. (2012). Hadoop the definitive guide . O'Reilly Media, Inc.

Williams, P., Sion, R., & Carbunar, B. (2008). Building castles out of
mud: practical access pattern privacy and correctness on untrusted
storage. ACM CCS, 139–148. doi:10.1145/1455770.1455790

Witten, H. I., & Frank, E. (2005). Data Mining: Practical machine

learning tools and techniques. Morgan Kaufmann.

Witten, Moffat, & Bell. (1999). Managing gigabytes: Compressing

and indexing documents and images. Morgan Kaufmann Publishing.

Witten, Moffat, & Bell. (1999). Managing Gigabytes: Compressing

and indexing documents and images. Morgan Kaufmann Publishing.

Wong, W. K., Cheung, D. W., Kao, B., & Mamoulis, N. (2009).

Secure knn computation on encrypted databases. Proc. of SIGMOD.
10.1145/1559845.1559862

Wong, W. K., Cheung, D. W., Kao, B., & Mamoulis, N. (2010).

Secure kNN computation on encrypted databases. Proc. ACM
SIGMOD Int. Conf. Manage. Data, 139-152.

Wu, G. (2008). Support vector machine classifier based on fuzzy

partition and neighborhood pairs. Jisuanji Yingyong , 28(1), 131–133.
doi:10.3724/SP.J.1087.2008.00131

Wu, H.-K., Lee, S. W.-Y., Chang, H.-Y., & Liang, J.-C. (2013,
March). Current status, opportunities and challenges of augmented
reality in education . Computers & Education , 62(C), 41–49.
doi:10.1016/j.compedu.2012.10.024

Wu, , Zhu, , & Wu, , & Ding. (2014). Data Mining with Big Data .
IEEE Transactions on Knowledge and Data Engineering , 26(1).

Xu W. Huang L. Fox A. Patterson D. Jordan M. I. (2009). Detecting

large-scale system problems by mining console logs. Proceedings of
the ACM 22nd Simposium on Operating Systems Principles
(SOSP’09), 117–132. 10.1145/1629575.1629587

Yang, L., Liu, & Yang. (2014). Secure dynamic searchable symmetric
encryption with constant document update cost. Proc.GLOBECOM.

Yang, Li, Liu, Yang, & M. (2014). Secure dynamic searchable

symmetric encryption with constant document update cost. In
Proceedings of GLOBCOM. IEEE.

Yang H. Xu Z. King I. Lyu M. (2010). Online learning for group

lasso. International Conference on Machine Learning (ICML’10).

Yi, X., Liu, F., Liu, J., & Jin, H. (2014). Building a network highway
for big data: Architecture and challenges . IEEE Network , 28(4), 5–
13. doi:10.1109/MNET.2014.6863125

Yu, J., Lu, P., Zhu, Y., Xue, G., & Li, M. (2013). Towards secure
multikeyword top-k retrieval over encrypted cloud data . IEEE
Transactions on Dependable and Secure Computing , 10(4), 239–250.
doi:10.1109/TDSC.2013.9

Zhang, E.-H. (2003). Research on paper currency recognition by

neural networks. In Machine Learning and Cybernetics, 2003
International Conference on (vol. 4). IEEE.

Zhang, B., & Zhang, F. (2011). An efficient public key encryption

with conjunctive-subset keywords search . Journal of Network and
Computer Applications , 34(1), 262–267.
doi:10.1016/j.jnca.2010.07.007

Zhang Q. Zhu Q. Zhani M. F. Boutaba R. (2012). Dynamic service

placement in geographically distributed clouds. Proc. IEEE 32nd Int’l
Conf. on Distributed Computing Systems (ICDCS’12), 526–535.
10.1109/ICDCS.2012.74

Zheng Z. Lyu M. R. (2009). A QoS-aware fault tolerant middleware

for dependable service composition. Proc. 39th Int’l Conf.
Dependable Systems and Networks (DSN’09), 239–248.
10.1109/DSN.2009.5270332
Zheng, Z., Ma, H., Lyu, M. R., & King, I. (2011). QoS-aware Web
service recommendation by collaborative filtering . IEEE Transactions
on Services Computing , 4(2), 140–152. doi:10.1109/TSC.2010.52

Zheng Z. Zhang Y. Lyu M. R. (2010). CloudRank: A QoS-driven

component ranking framework for cloud computing. Proc. Int’l Symp.
Reliable Distributed Systems (SRDS’10), 184–193.
10.1109/SRDS.2010.29

Zheng Z. Zhu J. Lyu M. R. (2013). Usage-created Big Data and Big

Data-as-a-Service: An Overview. IEEE International Congress on Big
Data.

Zhu J. Zheng Z. Zhou Y. Lyu M. R. (2013). Scaling service-oriented

applications into geodistributed clouds. Pro. IEEE Int’l Workshop on
Internet-based Virtual Computing Environment (iVCE’13).

Zikopoulos, P., & Eaton, C. (2011). Understanding big data: Analytics

for enterprise class hadoop and streaming data . McGraw-Hill
Osborne Media.
About the Contributors
Leonid Datta is a final year student of VIT University (Computer Science and
Engineering). Completed projects entitled “truncation technique for search
engine”, “Web design and data analysis” and “Universal Student Identification
using RFID Sensor”.

Sameera K. is a Research Associate, School of Computer Science and

Engineering, VIT University.

Chetan Kumar is a student of VIT University.

Venkatesan M. is an Assistant Professor in Department of Computing Science

and Engineering, National Institute of Technology Karnataka, Mangalore. He is
working in the area of big data analytics, data science and spatial data mining.
His research area includes databases, data warehouse, data mining, big data and
applications of data mining in various real-time applications. He has good
knowledge in data mining tools like Clementine and Rapidminer. He is the
Principal Investigator in ISRO funded project on data mining in a landslide. He
holds a BE, MTech and PhD in the area of spatial data mining.

Abhishek Mukherjee is a student of VIT University.

Prabhavathy P. is working as Associate Professor in School of Information

Technology and Enngineering, VIT University, Vellore. She has completed PhD
in the area of computational Intelligence. Her area of interest includes rough set,
fuzzy set, Sequential mining, data mining. She published number of paper in
international journals and international conference.

Pronay Peddiraju is a Computer Science Engineer with experience in game

engine technology, Augmented and Virtual Reality development and 3D object
visualisation. Pronay has performed research in fields of game technology and
problem solving.

Dhanasekaran Pillai received his Ph. D in Computer Science and Engineering

under the Faculty of Information and Communication Engineering from the
Anna University in 2015, India. In 2009 he was received M.E. in Computer
Science and Engineering from the Anna University and is currently the
Associate Professor at JGI-Jain College of Engineering, Belgaum, Karnataka,
India. He writes and presents widely on issues of data mining, information
extraction, and big data analytics, remote sensing, and is the member of Editorial
and reviewer board of IJRCIS.

Remya S. is now doing research in VIT University, Vellore. Completed M.Tech

in Calicut University during 2009-2011 and B.Tech in 2002-2006 from CUSAT.

Shweta Shinde is a VIT University student working in Genesys.

Ramasubbareddy Somula is now working as Assistant Professor in VIT

University. Done BTech in Alfa College of Engineering & Technology, Andhra
Pradesh. His MTech was in Srirama Engg College, AndhraPradesh.

Utkarsh Srivastava is a final year undergraduate from Computer Science

background studying in VIT University. He is extremely passionate about
learning and implementing technology. He also has deep interest in the field on
Bigdata. A polyglot at heart, he can code and communicate in several languages.