International Journal on Future Revolution in Computer Science & Communication Engineering ISSN: 2454-4248

Volume: 4 Issue: 1 385 – 387

_______________________________________________________________________________________________
Personalized Search Engine: A Review

Ms. Kajal Thakur, Prof. PragatiPatil

Department of Computer Science and Engineering

AbhaGaikwadPatil College of Engineering Mohgaon Nagpur Maharashtra India

Abstract: Now a days there is A major problem in mobile search is that the interactions between the users and search engines are
limited by the small form factors of the mobile devices. As a result, mobile users tend to submit shorter, hence, more ambiguous
queries compared to their web search counterparts. In order to return highly relevant results to the users, mobile search engines
must be able to profile the users’ interests and personalize the search results according to the user’s profiles. A personalized
mobile search engine (PMSE) that captures the users’ preferences in the form of concepts by mining their click through data. Due
to the importance of location information in mobile search, In this paper PMSE classifies these concepts into content concepts and
location concepts To characterize the diversity of the concepts associated with a query and their relevance’s to the user’s need,
four entropies are introduced to balance the weights between the content and location facets.

Keywords:--Clickthrough data, concept, location search, mobile search engine, ontology, profiling.

__________________________________________________*****_________________________________________________

I. INTRODUCTION Secondly, the existing Personalized Information Retrieval

System fails to analyze the search system using Smartphone
Social Network is a social structure made of individuals in Social Network environment. With an increasing number
called nodes, which are connected by one or more specific of web users using Smartphone and its individualized
types of interdependency, such as friendship, kinship, service under research, Smartphone environment does not
financial exchange, dislike, sexual relationships, or provide user’s search rankings suited to personal
relationships of beliefs, knowledge or prestige [1]. Social preferences. For example, when a user who wants to come
Network’s link represents not only the flow between by a pasta restaurant offering pasta for about 10$ and listens
personal information, but the relation status through to rock music asks for information search via Smartphone,
quantitative expression. The overall graph model of Social search results should also be prioritized and provided in
Network is composed of many nodes and the links that favor of user’s personalization taste. But, the existing
connect them, and each node’s direct/indirect connection systems do not show search rankings in consideration of
forms the entire network. However, the current Personalized individual’s tastes and tastes. Therefore, in this study
Systems based on Social Network were designed and differentiated search results are provided on the basis of
constructed under the PC and it didn’t provide the step by personalization information in User Profile Registry when a
step transferring methods from PC to Smartphone. To solve user carries out a search using Smartphone in Social
these problems, this research actively analyzes an Network environment. Finally this research attempts to
individual’s characteristic based on the Social Network correct uncertain or vague relation between users on the
environment and develops a Personalized Information existing Social Network environment and promote a more
Retrieval System which can search for what a user wants accurate and personalized information feeding, by
accurately. Personalized Information Retrieval System for suggesting a Personalized Information Retrieval System
efficient personalized information provision proposed in this using Social Network’s quantitative model..
study differs from existing ones in methodology as follow:
Firstly, as the system is built on the basis of NFC (Near field II. LITERATURE SURVEY
communication), it attempts to provide its own custom
service fast and easily using its information stored in NFC. Many existing personalized web search systems [2], [3], [4],
Once SNS and NFC Smartphone are associated with each [5] are based clickthrough data to determine users’
other, payment is made by touching a NFC tag when visiting preferences. Joachims [3] proposed to mine document
well known restaurants, and the information recorded in preferences from clickthrough data. Later, Ng et al. [4]
SNS is supposed to provide search results customized to proposed to combine a spying technique together with a
individual’s tastes and preferences when carrying out a novel voting procedure to determine user preferences.
search in individualized search system. That is, typing the Search queries can be classified as content (i.e., non-geo) or
same search keyword may bring different search results on location (i.e., geo) queries. Examples of location queries are
NFC Smartphone as individuals have different preferences. “hongkong hotels,” “museums in london,” and “virginia
historical sites.” In [6], Gan et al. developed a classifier to
385
IJFRCSCE | January 2018, Available @ http://www.ijfrcsce.org
_______________________________________________________________________________________
International Journal on Future Revolution in Computer Science & Communication Engineering ISSN: 2454-4248
Volume: 4 Issue: 1 385 – 387
_______________________________________________________________________________________________
classify geo and non-geo queries. It was found that a disadvantage of this architecture is that it brings high storage
significant number of queries were location queries focusing and computation costs when millions of users are using the
on location information. In order to handle the queries that search engine, and it also raises privacy concerns when
focus on location information, a number of location-based information about users is stored on the server. For client-
search systems designed for location queries have been side personalization, user information is collected and stored
proposed. Yokoji [8] proposed a location-based search on the client side (in the user’s computer or a
system for web documents. Location information was personalization agent), usually by installing a client software
extracted from the web documents, which was converted or plug-in on a user’s computer. In client side, not only the
into latitudelongitude pairs. When a user submits a query user’s search behavior but also his contextual activities (e.g.,
together with a latitude-longitude pair, the system creates a web pages viewed before) and personal information (e.g.,
search circle centered at the specified latitude-longitude pair emails, documents, and bookmarks) could be incorporated
and retrieves documents containing location information into the user profile [2].
within the search circle.
This allows the construction of a much richer user model for
More recently, Li et al. [9] proposed a probabilistic personalization. Privacy concerns are also reduced since the
topicbased framework for location- sensitive domain user profile is strictly stored and used on the client side.
information retrieval. Instead of modeling locations in Another benefit is that the overhead in computation and
latitude-longitude pairs, the model assumes that users can be storage for personalization can be distributed among the
interested in a set of locationsensitive topics. It recognizes clients. A main drawback of personalization on the client
the geographical influence distributions of topics, and side is that the personalization algorithm cannot use some
models it using probabilistic Gaussian Process classifiers. knowledge that is only available on the server side (e.g.,
PageRank score of a result document). Furthermore, due to
the limits of network bandwidth, the client can usually only
III. BASICS OF PERSONALIZED SEARCH process limited top results.

A. Creation of User Profile IV. CONTENT V/S HYPERLINK BASED

ANALYSIS
To provide personalized search results to users, personalized
A. Content Based Personalized Search By checking content
web search maintains a user profile for each individual. A
similarities between web pages and user profile personalized
user profile stores information about user interests and
search can be improved [3]. User’s interests can be
preferences. It is generated and updated by exploiting
automatically learned by classifying implicit user data.
userrelated information. Such information may include:
Search results are filtered or re-ranked by checking the
 Information about the user like age, gender, education, similarity of topics between search results and user profiles.
language, country, address, interest areas, and other User-issued queries and user-selected documents are
information. categorized into concept hierarchies that are accumulated to
generate a user profile. When the user issues a query, each
 Search history, including previous queries and clicked returned result is also classified. The documents are re-
documents. ranked based upon how well the document categories match
user interest profiles. Chirita et al. [4] use the ODP (Open
 Other user documents, such as bookmarks, favorite web
Directory Project, http://www.dmoz.org/) hierarchy to
sites, visited pages, and emails.
implement personalized search. User favorite topics nodes
B. Server-Side and Client-Side Implement Personalized web are manually specified in the ODP hierarchy. Each
search can be implemented on either server side (in the document is categorized into one or several topic nodes in
search engine) or client side (in the user’s computer or a the same ODP hierarchy. The distances between the user
personalization agent) topic nodes and the document topic nodes are then used to
re-rank search results.
[1]. For server-side personalization, user profiles are built,
updated, and stored on the search engine side. User B. Hyperlink Based Personalized Search Hyperlink Analysis
information is directly incorporated into the ranking process, significantly improves the relevance of the web search
or is used to help process initial search results. The results so that all major search engines claim to use some
advantage of this architecture is that the search engine can type of hyperlink analysis. Web information retrieval mainly
use all of its resources, for example link structure of the focuses on hyperlink structures of the Web, like with Web
whole web, in its personalization algorithm. Also, the search engine Google. In personalized Web searches, the
personalization algorithm can be easily adapted without any hyperlink structures of the Web are also becoming
client efforts. This architecture is adopted by some general important. The use of personalized PageRank to enable
search engines such as Google Personalized Search. The personalized Web searches was first proposed in [6], where
it was suggested as a modification of the global PageRank
386
IJFRCSCE | January 2018, Available @ http://www.ijfrcsce.org
_______________________________________________________________________________________
International Journal on Future Revolution in Computer Science & Communication Engineering ISSN: 2454-4248
Volume: 4 Issue: 1 385 – 387
_______________________________________________________________________________________________
algorithm, which computes a universal notion of importance could improve search accuracy for some queries, but they
of a Web page. The computation of (personalized) harmed performance for more queries. Personalized Web
PageRank scores was not addressed beyond the original search yields significant improvements over generic Web
algorithm. Experiments[7] concluded that the use of search for queries with high click entropy. For the queries
personalized PageRank scores can improve a Web search. with low click entropy, personalization methods performed
Crawling (process of gathering the web pages by the search similarly or even worse than generic search. As personalized
engine) and ranking are the main uses of hyperlink analysis. search had different effectiveness for different kinds of
In this approach, web crawler which is a software program queries, queries should not be handled in the same manner
to browse WWW in automated methodical manner find with regard to personalization. No personalization
more and more web pages linked to the source page with the algorithms can outperform others for all queries. Different
assumption of nearly all the linked web pages are on same methods have different strengths and weaknesses.
topic. This process repeats for each set of web pages until no
more linked pages. Then crawler of the search engine orders REFERENCES
the web pages by the quality. To judge high quality pages, [1] Ji-Rong Wen, Zhicheng Dou, Ruihus Song, “Personalized
hyperlink analysis is used. In this case, search engine Web Search”, Microsoft Research Asia, Beijing, China,
assumes that the source page pointed to many pages is of 2009
higher quality than a source page pointed to few numbers of [2] P.A. Chirita, C. Firan, and W. Nejdl, “Summarizing Local
pages. For this ranking process, hyperlink analysis involves Context to Personalize Global Web Search”, Proc. ACM
with connectivitybased ranking, PageRank and HITS Int’l Conf. Infor. and Knowledge Management (CIKM),
2006.
(Hyperlink-Induced Topic Search) algorithms etc. In
[3] Zhongming Ma, Gautam Pant and Olivia R. Liu Sheng.
addition to produce a quality and relevant web results,
“Interest-Based Personalized Search,” ACM Transactions
hyperlink analysis have several advantages like finding on Information Systems, Vol 25, Issue 2, Article 5,
mirrored hosts, web page categorization and identify the February 2007.
geographical scope of the search etc. But in this approach, [4] Chirita P.A., Nejdl W., Paiu R., and Kohlschu¨tter C.
search engine has to deal with more details consist even “Using ODP metadata to personalize search”. In Proc. 31st
with unnecessary stuffs also. It becomes wastage of the Annual Int. ACM SIGIR Conf. on Research and
resources. Development in Information Retrieval, 2005, pp. 178–185.
[5] Aktas M. S., M. A. Nacar and F. Menczer, 2004. “Using
III.CONCLUSION hyperlink features to personalize web search”. Adv. Web
Mining Web Usage Anal., 3932: 104-115.
There are number of researches which are conducting to [6] Page L., Brin S., Motwani R., and Winograd T. “The
minimize some of the drawbacks of Personalized Web Pagerank Citation Ranking: bringing order to the Web”,
Search. Hyperlink based personalization algorithms work Technical report, Stanford Dig. Lib. Tech. Project, 1998.
only for repeated queries, they are simple and stable. The [7] K. Sugiyama, K. Hatano, and M. Yoshikawa. “Adaptive
topical interestbased personalized search algorithms web search based on user profile constructed without any
effort from users”. In Proceedings of the 13th international
implemented were not as stable as the click-based. They
conference on World Wide Web, pages 675–684, 2004.

387
IJFRCSCE | January 2018, Available @ http://www.ijfrcsce.org
_______________________________________________________________________________________