![Jan Zizka (Eds) : CCSIT, SIPP, AISC, PDCTA - 2013
pp. 413–420, 2013. © CS & IT-CSCP 2013 DOI : 10.5121/csit.2013.3647
A NOVEL APPROACH FOR SELECTION OF
LEARNING OBJECTS FOR PERSONALIZED
DELIVERY OF E-LEARNING CONTENT
Mrs. D. Anitha1
, DR. C. Deisy2
1
Assistant Professor, 2
Associate Professor,
Thiagarajar College of Engineering, Madurai, Tamilnadu, India 625006
anithad@tce.edu1
,cdcse@tce.edu2
ABSTRACT
Personalized E-learning, as an intelligent package of technology enhanced education tends to
overrule the traditional practices of static web based E-learning systems. Delivering suitable
learning objects according to the learners’ knowledge, preferences and learning styles makes
up the personalized E-learning. This paper proposes a novel approach for classifying and
selecting learning objects for different learning styles proposed by Felder and Silverman The
methodology adheres to the IEEE LOM standard and maps the IEEE LO Metadata to the
identified learning styles based on rule based classification of learning objects. A pilot study on
the research work is performed and evaluation of the system gives an encouraging result.
KEYWORDS
E-learning; Learning Styles; IEEE LOM, Classification;
1. INTRODUCTION
E-learning, regarded as technology package of education has attained significance due to its
advantages of learning at any time, any pace and from anywhere. Learning is always a cognitive
activity which differs from learner to learner and so raises the need of personalized E-learning.
Recent research on the learning process has shown that there is a difference of learning styles
among the individual learners and different teaching and learning resources are indispensable to
satisfy their learning needs.
Learning Object (LO), a collection of content items, practice items and assessment items for
education is an inevitable constituent of any E-learning systems. LOs’ metadata need to be
recorded in order to classify them according to the style of the learner. The IEEE Learning
Technology Standards Committee (LTSC) has prescribed a standard for describing the metadata
instance for an LO to facilitate search, evaluation, acquisition, and use of learning objects
[2][6][8].
This paper proposes a novel methodology which classifies any LO that adheres to the IEEE LOM
standard into different teaching strategies, then mapping these strategies to different learning
styles proposed by Felder Silverman catering to learner needs.](https://image.slidesharecdn.com/csit3647-180201095350/85/A-NOVEL-APPROACH-FOR-SELECTION-OF-LEARNING-OBJECTS-FOR-PERSONALIZED-DELIVERY-OF-E-LEARNING-CONTENT-1-320.jpg)
![414 Computer Science & Information Technology (CS & IT)
2. RELATED WORKS
Deriving a personalized E-learning environment is a crucial area of research now. Recent
research works [1][4][6][7][8][10][11] illustrate the need of personalization of E-learning
systems. Adaptation of Learning objects (LOs) is considered as one of the major aspects of
personalization [1][4][6][7][11]. The need of customized LO repositories and selection is
specified in [8][18]. LO selection problem in intelligent learning systems is addressed in [6]
which produce a decision model not compliant with IEEE LOM standard. An architecture based
on Semantic Web for E-learning is defined in [1] which understands the learners’ preferences and
interpreting it as ontology. The use of learning content in different contexts and in different
formats most appropriate for an individual learner is presented in [10].
There exists different methodologies to assess learner styles and among those Felder Silverman
model plays a significant role [3][9][12][13]. A research work [12] states a way for generation of
personalized courses from suitable repositories of learning nodes. A theoretical work is proposed
[7] for the definition of learning profiles and classification of the student within a given learning
profile. Another work proposed [15] a learning style classification mechanism with k-nearest and
genetic algorithm to classify and then identify the learners’ learning styles. The study did not
refer any proven learning styles. A study [19] used Bloom’s taxonomy and Genetic algorithms to
personalize e-learning. Franzoni [5] suggested the teaching strategies according to Felder
Silverman model and appropriate electronic media.
3. FELDER SILVERMAN LEARNING STYLE
Learning styles are various approaches or ways of learning. Though there are many learning
models available, such as David Kolb’s, Pask’s, Honey & Mumford’s, Gregorc’s etc., Felder and
Silverman Learning style model has been taken frequently by many research
works[5][7][12][16][17]. This model classifies students’ preferred learning style on four major
dimensions [3][9][13] according to their responses for an Interactive Learning Style(ILS)
Questionnaire. Table1 shows the four dimensions and two different behaviors of each dimension.
This proposed research work uses this model because its ILS Questionnaire gives us the
possibility of linking directly its results to automatic adaptive environments.
Table 1. Felder Silverman Model of Learning Style
Dimension Learning Style Explanation
Participation Active (A) Needs hands on work experiments
Reflective(R) Passive and prefer to think things
Processing Sensory (S) Believes Concrete facts
Intuitive (I) Conceptual and theoretical view
Presentation Visual (Vi) Prefers diagrams, pictures, visual presentations
Verbal (V) Audio Narration or display of text
Organization Sequential(Sq) Needs information in a linear fashion
Global (G) Prefers overall view
From Table1, it is inferred there are 16 different combinations of the learner characteristics for
eg.{(A,S,Vi,Sq)-1, (A,S,Vi,G)-2,… (R,I,V,Sq)-15, (R,I,V,G)-16}](https://image.slidesharecdn.com/csit3647-180201095350/85/A-NOVEL-APPROACH-FOR-SELECTION-OF-LEARNING-OBJECTS-FOR-PERSONALIZED-DELIVERY-OF-E-LEARNING-CONTENT-2-320.jpg)
![416 Computer Science & Information Technology (CS & IT)
Table 2. Recommended Teaching Strategies
Dimension Learning
Behaviour
Recommended Teaching strategies
Processing Active (A) Interactive Demonstrations, Exercise Problems
Reflective(R) Visual Presentation, Examples
Presentation Sensory (S) Examples, Case Study
Intuitive (I) Theoretical Explanations, Visual Presentations,
Simulations
Participation Visual (Vi) Visual Presentations, Visual Overview
Verbal (V) Theoretical explanations, Verbal overview
Organization Sequential(Sq) Visual Presentations, Theoretical explanations
Global (G) Visual or verbal overview
The LOs are maintained in XML representation following IEEE LOM standard. It is observed
that the classification of the LOs could be made based on metadata elements [5][8][10]. The
proposed work takes 6 metadata elements listed as: Structure (1.7), Aggregation Level (1.8),
Technical Format (4.1), Interactivity Type (5.1), Learning Resource Type (5.2), and
Interactivity Level (5.3). Based on the values of the metadata elements of the LOs they are
classified into any one of the 8 teaching strategies. LOs are mapped to the respective teaching
strategies using rule based classification and Table 3 gives the rule based classification strategy
on the metadata elements and their permissible values.
Table 3. Rule based strategy for classifying Learning Objects
Teaching
Strategy
Classification
IEEE LO Metadata Elements & values
1.7 1.
8
4.1 5.1 5.2 5.3l
Topic
Overview :
Visual(OV)
Collection
/
Hierarchic
al
2 Video/Mpeg Expositive Index Low
Topic
Overview :
Verbal (OVe)
Collection
/
Hierarchic
2 Text/Html Expositive Index Low
Theoretical
explanations
(T)
Atomic 1 Text/Html Expositive Narrative
text/ Lecture
Low
Visual
Presentations
(VP)
Atomic 1 Video/Mpeg Mixed Diagram,
Figure,
Graph, Table
Med
Interactive
Demonstrations
(ID)
Atomic 1 Application Active Questionnaire Very
high
Examples (E) Atomic 1 Text/Html Expositive Experiment Low
Simulations (S) Atomic 1 Application
Video/Mpeg
Expositive Simulation Low
Exercise
Problems(EP)
Atomic 1 Text/Html Active Self
assessment
High
Case Study(C) Atomic 1 Text/Html Expositive Problem
Statement
Low](https://image.slidesharecdn.com/csit3647-180201095350/85/A-NOVEL-APPROACH-FOR-SELECTION-OF-LEARNING-OBJECTS-FOR-PERSONALIZED-DELIVERY-OF-E-LEARNING-CONTENT-4-320.jpg)
![Computer Science & Information Technology (CS & IT) 419
of an LO of teaching strategy t by a particular set of learners whether t is recommended to them
or not.
Figure 2 Rating of the recommended teaching strategies
6. CONCLUSION AND FUTURE WORKS
This research work is an experimental effort for approaching personalized E-learning with respect
to differing learning styles. It prescribes a way of mapping different learner styles with suitable
learning objects. It also provides a means for accessing any learning objects which adheres to the
IEEE LOM standard in public repositories to be classified with its metadata information.. The
study does not declare the system as a complete alternate to the human teaching system. But it
could be a supplementary process in selective topics.
The future work pertains to the extension of the system into a complete prototype which
considers the learner knowledge level, specific interests in addition to the learning style. Also, the
system would be made to change its decision on knowledge level and learning style dynamically
and also considering the learners’ interests, time spent with each topic and additional information
accessed during navigation.
REFERENCES
[1] Igor Keleberda, Victoria Repka, Yevgen Biletskiy, “Semantic Mining Based on the Learner’s
Preferences”, Proc. Of IEEE Canadian Conference on Electrical and Computer Engineering, p. 502-
504, May 2006.
[2] IEEE Learning Technologies Committee, “Draft Standarad for Learning Object Metadata”, July 2002.
[3] R.M. Felder and L.K. Silverman, "Learning and Teaching Styles in Engineering Education," Journal
of Engineering Education, Vol. 78, No.7, p. 674-681, 1988.
[4] Idris, N. Yusof, N. Saad, P., “Concept-Based Classification for Adaptive Course Sequencing Using
Artificial Neural Network”, Proc. of Ninth International IEEE Conference on Intelligent Systems
Design and Applications, p. 956-960, 2009.
[5] Franzoni, A. L., & Assar, S., “Student Learning Styles Adaptation Method Based on Teaching
Strategies and Electronic Media”, Educational Technology & Society, Vol.12, No. 4, p. 15–29, 2009.
[6] Phytogoras Karampiperis, Demetrios Sampson, “Adaptive Learning Object Selection in Intelligent
learning systems”, Journal of Interactive Learning Research, Special issue on computational
Intelligence in Web-based Education, Vol. 15, No.4,p.389-407,Nov 2004.
[7] Luciana A M Zaina, Graça Bressan , “Classification of Learning Profile Based on Categories of
Student Preferences”, 38th ASEE/IEEE Frontiers in Education Conference, pp. F4E-1 - F4E-6 ,
October 2008, Saratoga Springs, NY.
[8] Peter Brusilovsky,Julita Vassileva, “Course sequencing techniques for large-scale webbased
education”, Int. J. Cont. Engineering Education and Lifelong Learning, Vol. 13, Nos.1/2, 2003, p.75-
94.](https://image.slidesharecdn.com/csit3647-180201095350/85/A-NOVEL-APPROACH-FOR-SELECTION-OF-LEARNING-OBJECTS-FOR-PERSONALIZED-DELIVERY-OF-E-LEARNING-CONTENT-7-320.jpg)
![420 Computer Science & Information Technology (CS & IT)
[9] Zywno, M. ,”A Contribution to Validation of Score Meaning for Felder-Soloman's Index of Learning
Styles”, ASEE Conference, Nashville, Tennessee,2003.
[10] Korneliya Yordanova, “Meta-Data Application in Development, Exchange and Delivery of Digital
Reusable Learning Content”, Interdisciplinary Journal of Knowledge and Learning Objects, Volume
3, p. 229-237, 2007.
[11] C. Limongelli, F. Sciarrone, G. Vaste, “LS-Plan : An Effective Combination of Dynamic Courseware
Generation and Learning Styles in Web-Based Education”, Proc. Fifth International Conference in
Adaptive Hypermedia and Adaptive Web-Based Systems, p. 133-142, 2008.
[12] C. Limongelli, F. Sciarrone, G. Vaste, “An Application of the LS-Plan System to an Educational
Hypermedia”, Intl J Web-Based Learning and Teaching Technologies, vol.4. no.1, p. 15-34, 2009.
[13] S Viola, S. R., Graf, S., Kinshuk, & Leo, T.,” Investigating relationships,within the Index of
Learning Styles: A data-driven approach”, International Journal of Interactive Technology and Smart
Education, Vol. 4, No.1, p. 7–18, 2007.
[14] Huey-Ing Liu Min-Num Yang, “QoL guaranteed adaptation and personalization in E-learning
systems”, Transactions on Education, Vol. 48, Issue4, p. 676 – 687, 2005.
[15] Yi-Chun Chang et al., “A learning style classification mechanism for e-learning”, Computers &
Education, Vol.53, p. 273-285,2009.
[16] C.A. Carver, R.A. Howard, W.D. Lane, “Enhancing Student Learning Through Hypermedia
Courseware and Incorporation of Student Learning Styles,” Transactions on Education, Vol.42, No.1,
p. 33-38, February 1999.
[17] E. Sanigneto, N. Capuano, M. Gaeta, A. Micarelli, “Adaptive Course Generation through Learning
Styles Representation”, Universal Access Information Society(UAIS’08), Vol.7,No.1/2, p.1-23, 2008.
[18] Karampiperis, P., & Sampson, D. (2005). Adaptive Learning Resources Sequencing in Educational
Hypermedia Systems”, Educational Technology & Society, Vol. 8, No.4, p. 128-147.
[19] Anitha D, Deisy C, “Deriving a prototype for the dynamic generation of learning path in an e-learning
environment using Genetic algorithm”, International Journal of Innovation and Learning, Article in
Press.
Authors
Anitha D is a research scholar of Anna University of Technology, Madurai, India and
currently working as Assistant Professor in the Department of Computer Applications,
Velammal College of Engineering & Technology,Madurai, India. She has done her
Masters in Computer Applications and is interested in the application of Knowledge
based techniques in the field of E-learning. She firmly believes that improving e-learning
will enhance the teaching-learning process thereby improving the potential of excellence
in educational institutions.
Dr. C. Deisy received the Ph.D. degree in Computer Science from Anna University,
Chennai in 2010. She is currently working as Associate Professor in the Department of
Computer Science and Engineering, Thiagarajar College of Engineering ,Madurai,
India. She has authored numerous scientific publications. Her main research interests are
in the field of Data Mining and its applications. She has published her research works in
many International Journals and conferences. She is the Journal Reviewer of Elsevier
Journal of Applied Soft Computing since 2010.](https://image.slidesharecdn.com/csit3647-180201095350/85/A-NOVEL-APPROACH-FOR-SELECTION-OF-LEARNING-OBJECTS-FOR-PERSONALIZED-DELIVERY-OF-E-LEARNING-CONTENT-8-320.jpg)
A NOVEL APPROACH FOR SELECTION OF LEARNING OBJECTS FOR PERSONALIZED DELIVERY OF E-LEARNING CONTENT
- 1. Jan Zizka (Eds) : CCSIT, SIPP, AISC, PDCTA - 2013 pp. 413–420, 2013. © CS & IT-CSCP 2013 DOI : 10.5121/csit.2013.3647 A NOVEL APPROACH FOR SELECTION OF LEARNING OBJECTS FOR PERSONALIZED DELIVERY OF E-LEARNING CONTENT Mrs. D. Anitha1 , DR. C. Deisy2 1 Assistant Professor, 2 Associate Professor, Thiagarajar College of Engineering, Madurai, Tamilnadu, India 625006 [email protected] ,[email protected] ABSTRACT Personalized E-learning, as an intelligent package of technology enhanced education tends to overrule the traditional practices of static web based E-learning systems. Delivering suitable learning objects according to the learners’ knowledge, preferences and learning styles makes up the personalized E-learning. This paper proposes a novel approach for classifying and selecting learning objects for different learning styles proposed by Felder and Silverman The methodology adheres to the IEEE LOM standard and maps the IEEE LO Metadata to the identified learning styles based on rule based classification of learning objects. A pilot study on the research work is performed and evaluation of the system gives an encouraging result. KEYWORDS E-learning; Learning Styles; IEEE LOM, Classification; 1. INTRODUCTION E-learning, regarded as technology package of education has attained significance due to its advantages of learning at any time, any pace and from anywhere. Learning is always a cognitive activity which differs from learner to learner and so raises the need of personalized E-learning. Recent research on the learning process has shown that there is a difference of learning styles among the individual learners and different teaching and learning resources are indispensable to satisfy their learning needs. Learning Object (LO), a collection of content items, practice items and assessment items for education is an inevitable constituent of any E-learning systems. LOs’ metadata need to be recorded in order to classify them according to the style of the learner. The IEEE Learning Technology Standards Committee (LTSC) has prescribed a standard for describing the metadata instance for an LO to facilitate search, evaluation, acquisition, and use of learning objects [2][6][8]. This paper proposes a novel methodology which classifies any LO that adheres to the IEEE LOM standard into different teaching strategies, then mapping these strategies to different learning styles proposed by Felder Silverman catering to learner needs.
- 2. 414 Computer Science & Information Technology (CS & IT) 2. RELATED WORKS Deriving a personalized E-learning environment is a crucial area of research now. Recent research works [1][4][6][7][8][10][11] illustrate the need of personalization of E-learning systems. Adaptation of Learning objects (LOs) is considered as one of the major aspects of personalization [1][4][6][7][11]. The need of customized LO repositories and selection is specified in [8][18]. LO selection problem in intelligent learning systems is addressed in [6] which produce a decision model not compliant with IEEE LOM standard. An architecture based on Semantic Web for E-learning is defined in [1] which understands the learners’ preferences and interpreting it as ontology. The use of learning content in different contexts and in different formats most appropriate for an individual learner is presented in [10]. There exists different methodologies to assess learner styles and among those Felder Silverman model plays a significant role [3][9][12][13]. A research work [12] states a way for generation of personalized courses from suitable repositories of learning nodes. A theoretical work is proposed [7] for the definition of learning profiles and classification of the student within a given learning profile. Another work proposed [15] a learning style classification mechanism with k-nearest and genetic algorithm to classify and then identify the learners’ learning styles. The study did not refer any proven learning styles. A study [19] used Bloom’s taxonomy and Genetic algorithms to personalize e-learning. Franzoni [5] suggested the teaching strategies according to Felder Silverman model and appropriate electronic media. 3. FELDER SILVERMAN LEARNING STYLE Learning styles are various approaches or ways of learning. Though there are many learning models available, such as David Kolb’s, Pask’s, Honey & Mumford’s, Gregorc’s etc., Felder and Silverman Learning style model has been taken frequently by many research works[5][7][12][16][17]. This model classifies students’ preferred learning style on four major dimensions [3][9][13] according to their responses for an Interactive Learning Style(ILS) Questionnaire. Table1 shows the four dimensions and two different behaviors of each dimension. This proposed research work uses this model because its ILS Questionnaire gives us the possibility of linking directly its results to automatic adaptive environments. Table 1. Felder Silverman Model of Learning Style Dimension Learning Style Explanation Participation Active (A) Needs hands on work experiments Reflective(R) Passive and prefer to think things Processing Sensory (S) Believes Concrete facts Intuitive (I) Conceptual and theoretical view Presentation Visual (Vi) Prefers diagrams, pictures, visual presentations Verbal (V) Audio Narration or display of text Organization Sequential(Sq) Needs information in a linear fashion Global (G) Prefers overall view From Table1, it is inferred there are 16 different combinations of the learner characteristics for eg.{(A,S,Vi,Sq)-1, (A,S,Vi,G)-2,… (R,I,V,Sq)-15, (R,I,V,G)-16}
- 3. Computer Science & Information Technology (CS & IT) 415 4. PROPOSED METHODOLOGY The proposed system consists of two important modules – Learner Profile modeling module, Pedagogical module. The system architecture is given in Figure 1. Figure 1 Proposed System Architecture 4.1. Learner Profile Modeling Initially, the learners are asked to answer Felder Silverman Interactive Learning Style Questionnaire and their preferred style of learning is recorded in Learner Profile Database. Learner Profile model studies the learner and updates his metadata on the preferred learning style. The learner is classified into one of the 16 categories listed above. The learner profile is represented in XML. 4.2. Learning Object (LO) Repository & Classified Los Each learning objective of the course is designed in multi ways representing the different learner characteristics and stored in LO Repository. The identified learning styles according to the Felder Silverman model are given in Table 1. For a learning objective of the unit, the LOs are designed with the following strategies: • Unit Overview – Visual(OV) • Unit Overview Verbal (OVe) • Theoretical explanations (T) – Lecture materials in the form of narrative text • Visual Presentations (VP) – Lecture materials along with supporting visuals • Interactive Demonstrations(ID) – Prompting learners to participate and give answers • Examples (E) – Illustrations, instances E.g. Pointer increment • Simulations (S) – Small models of the systems E.g. Navigation of arrays • Exercise Problems(EP) – Prompting learner to complete small problems • Case study(C) - A complete explanation of the existing system Also, the LOs are given a suitability rating (SLOkij) which decides the suitability of the learning object j in the topic i for a learner with learning style j. SLOkij = 0 or 1 (Initially 1) (1) k = 1..16 reflecting 16 dimensions of Felder –Silverman Learning Style i = 1..n where n- number of topics in a unit of syllabus j = 1..m where m – number of LOs for a particular topic i Table 2 shows the recommended teaching strategies for different learning behavior.
- 4. 416 Computer Science & Information Technology (CS & IT) Table 2. Recommended Teaching Strategies Dimension Learning Behaviour Recommended Teaching strategies Processing Active (A) Interactive Demonstrations, Exercise Problems Reflective(R) Visual Presentation, Examples Presentation Sensory (S) Examples, Case Study Intuitive (I) Theoretical Explanations, Visual Presentations, Simulations Participation Visual (Vi) Visual Presentations, Visual Overview Verbal (V) Theoretical explanations, Verbal overview Organization Sequential(Sq) Visual Presentations, Theoretical explanations Global (G) Visual or verbal overview The LOs are maintained in XML representation following IEEE LOM standard. It is observed that the classification of the LOs could be made based on metadata elements [5][8][10]. The proposed work takes 6 metadata elements listed as: Structure (1.7), Aggregation Level (1.8), Technical Format (4.1), Interactivity Type (5.1), Learning Resource Type (5.2), and Interactivity Level (5.3). Based on the values of the metadata elements of the LOs they are classified into any one of the 8 teaching strategies. LOs are mapped to the respective teaching strategies using rule based classification and Table 3 gives the rule based classification strategy on the metadata elements and their permissible values. Table 3. Rule based strategy for classifying Learning Objects Teaching Strategy Classification IEEE LO Metadata Elements & values 1.7 1. 8 4.1 5.1 5.2 5.3l Topic Overview : Visual(OV) Collection / Hierarchic al 2 Video/Mpeg Expositive Index Low Topic Overview : Verbal (OVe) Collection / Hierarchic 2 Text/Html Expositive Index Low Theoretical explanations (T) Atomic 1 Text/Html Expositive Narrative text/ Lecture Low Visual Presentations (VP) Atomic 1 Video/Mpeg Mixed Diagram, Figure, Graph, Table Med Interactive Demonstrations (ID) Atomic 1 Application Active Questionnaire Very high Examples (E) Atomic 1 Text/Html Expositive Experiment Low Simulations (S) Atomic 1 Application Video/Mpeg Expositive Simulation Low Exercise Problems(EP) Atomic 1 Text/Html Active Self assessment High Case Study(C) Atomic 1 Text/Html Expositive Problem Statement Low
- 5. Computer Science & Information Technology (CS & IT) 417 LO classification involves the use of Java code to parse the XML metadata elements of LOs and classify them into one of the teaching strategies. Eg. A. If Value (5.3) = “Medium” Then LO class = “VP” (Visual Presentations) B. If Value (1.8) = “2” and Value (4.1) = “Video/Mpeg” Then LO class = “OVe” 4.3. Pedagogical Module Pedagogical Module comprises two sub modules: LO Selector, Instructional Planner. LO Selector selects all the LOs for a given topic from the metadata element 1.5-Keyword which describes the topic of the Learning Object. Instructional Planner matches the Learner Profile model with the selected LOs and performs classification and a teaching sequence for a particular topic. Table 4 shows the different learning dimensions of Felder & Silverman and their matching teaching strategies. Table 4 Learning Styles and their matching teaching strategies SUITABILITY OF THE TEACHING STRATEGIES OV Ove T VP ID E S EP C Learning Style Active NA NA - - X - - X - Reflexive NA NA X X - X X - X Sensory NA NA NA NA NA X - - X Intuitive NA NA NA NA NA - X X - Visual X - - X NA NA NA NA NA Verbal - X X - NA NA NA NA NA Sequential - - NA NA NA NA NA NA NA Global X X NA NA NA NA NA NA NA X suitable, - Not suitable, NA Not applicable for this category Based on the teaching/learning strategies prescribed in Table4, a combination of the teaching/learning strategies is provided by the Instructional Planner: for eg. Teaching strategies for a category (A,S,Vi,Sq) : ID,E,C,VP ; (R,I,Ve,G) : T,,S,EP,T,OVe Though the pedagogical module recommends the given teaching sequence, it is not enforced on the learner to follow the recommended teaching strategy due to their behavioral changes. The learners are given the choice of following the recommended sequence or navigating between selective LOs of a particular topic in his/her preferred order. 5. IMPLEMENTATION AND EVALUATION OF THE SYSTEM The proposed system was implemented and tested for a unit of the PG Course “Programming in C”: Arrays & pointers in C. Learning Objects were designed according to the defined teaching strategies and LO repository is created with 72 LOs (8 topics with 9 teaching strategies for each). A web based application was developed with PHP and MYSQL running with Apache Tomcat Web Server where Java Code was used to implement Learner profile modeling and pedagogical module. A set of 99 PG students were asked to undertake the pilot study of the research work. The students were initially asked to respond to the ILS Questionnaire giving a reasonable amount of time and their profile is recorded. The students were asked to undertake the course by going through all the LOs of a particular topic (i.e.) not restricting them to the recommended teaching
- 6. 418 Computer Science & Information Technology (CS & IT) sequence. At the end of the delivery of each LO, a short feedback questionnaire on the usefulness of the LO is prompted to the learner. The LO is rated with the user feedback having positive added values for agreement(Strongly agree : 2; Agree : 1) and negative values for disagreement (Disagree : -1 ; Strongly disagree : -2). Let the feedback value given by a student of learning style k for a single LO is Valij where i is the topic number and j is the LO number where Valij is the feedback value between 12 and -12. Applying min-max normalization, the normalized feedback NVal ij is obtained from the given formula: NVal ij = (Val ij + 12) / 24 ; range: 0 to 1 (2) This normalized feedback given by the student of learning style k is used to dynamically adjust the suitability factor (SLO) of an LO to a particular learning style. Lesser the feedback value, lesser the suitability of the system is the hypothesis chosen which is universally accepted for evaluating any feedback. The suitability factor is obtained from the given formula which depreciates the suitability factor from 1 to 0 by its uselessness. SLOkij = SLOkij - ( (1- NVal ij) / n ) (3) where n = number of students of style k The suitability factor of an LO is thus adjusted for n students of learning style k. After evaluating n students of learning style k, the LO is accepted as a part of recommended teaching strategy for the learning style k if SLOkij value for the LO is above a given threshold, in this case, threshold is chosen as 0.5. Table 7 shows the mean cumulative suitability factors of all LOs in the given topics for all the existing learning styles. From Table 7, the following observations are made. For a particular teaching strategy t (e.g. T – Narrative text), two sets of learners are formed. The first set of learners (t1) is with the learning style which has t in its recommended teaching sequence and the second set of learners (t2) does not have t in their recommended teaching sequence. The mean value of suitability factors of both the sets are calculated and represented in graph as Figure 2. Table 7 Mean of Cumulative Suitability factors of LOs for 11 learning styles Learning Styles 1 2 4 5 6 8 9 12 13 14 16 Teaching Strategies OV 0.4 0.7 0.4 0.3 0.8 0.4 0.5 0.4 0.4 0.9 0.4 Ove 0.2 0.3 0.8 0.5 0.3 0.9 0.3 0.8 0.4 0.4 0.8 T 0.2 0.2 0.9 0.2 0.3 0.8 0.3 0.9 0.3 0.4 0.9 VP 0.8 0.7 0.4 0.8 0.8 0.3 0.8 0.4 0.9 0.9 0.4 ID 0.7 0.8 0.7 0.8 0.8 0.8 0.4 0.3 0.5 0.5 0.4 E 0.8 0.8 0.8 0.4 0.5 0.5 0.8 0.9 0.9 0.9 0.8 S 0.4 0.3 0.3 0.7 0.7 0.7 0.8 0.7 0.8 0.7 0.6 EP 0.8 0.7 0.7 0.8 0.7 0.8 0.4 0.5 0.8 0.8 0.8 C 0.7 0.6 0.7 0.3 0.4 0.3 0.8 0.8 0.7 0.7 0.8 The observation of the graph clearly shows that their preference to the recommended teaching strategy is always higher to the non recommended teaching strategy. And also the standard deviation values of the suitability factors states that there is no significant difference in the rating
- 7. Computer Science & Information Technology (CS & IT) 419 of an LO of teaching strategy t by a particular set of learners whether t is recommended to them or not. Figure 2 Rating of the recommended teaching strategies 6. CONCLUSION AND FUTURE WORKS This research work is an experimental effort for approaching personalized E-learning with respect to differing learning styles. It prescribes a way of mapping different learner styles with suitable learning objects. It also provides a means for accessing any learning objects which adheres to the IEEE LOM standard in public repositories to be classified with its metadata information.. The study does not declare the system as a complete alternate to the human teaching system. But it could be a supplementary process in selective topics. The future work pertains to the extension of the system into a complete prototype which considers the learner knowledge level, specific interests in addition to the learning style. Also, the system would be made to change its decision on knowledge level and learning style dynamically and also considering the learners’ interests, time spent with each topic and additional information accessed during navigation. REFERENCES [1] Igor Keleberda, Victoria Repka, Yevgen Biletskiy, “Semantic Mining Based on the Learner’s Preferences”, Proc. Of IEEE Canadian Conference on Electrical and Computer Engineering, p. 502- 504, May 2006. [2] IEEE Learning Technologies Committee, “Draft Standarad for Learning Object Metadata”, July 2002. [3] R.M. Felder and L.K. Silverman, "Learning and Teaching Styles in Engineering Education," Journal of Engineering Education, Vol. 78, No.7, p. 674-681, 1988. [4] Idris, N. Yusof, N. Saad, P., “Concept-Based Classification for Adaptive Course Sequencing Using Artificial Neural Network”, Proc. of Ninth International IEEE Conference on Intelligent Systems Design and Applications, p. 956-960, 2009. [5] Franzoni, A. L., & Assar, S., “Student Learning Styles Adaptation Method Based on Teaching Strategies and Electronic Media”, Educational Technology & Society, Vol.12, No. 4, p. 15–29, 2009. [6] Phytogoras Karampiperis, Demetrios Sampson, “Adaptive Learning Object Selection in Intelligent learning systems”, Journal of Interactive Learning Research, Special issue on computational Intelligence in Web-based Education, Vol. 15, No.4,p.389-407,Nov 2004. [7] Luciana A M Zaina, Graça Bressan , “Classification of Learning Profile Based on Categories of Student Preferences”, 38th ASEE/IEEE Frontiers in Education Conference, pp. F4E-1 - F4E-6 , October 2008, Saratoga Springs, NY. [8] Peter Brusilovsky,Julita Vassileva, “Course sequencing techniques for large-scale webbased education”, Int. J. Cont. Engineering Education and Lifelong Learning, Vol. 13, Nos.1/2, 2003, p.75- 94.
- 8. 420 Computer Science & Information Technology (CS & IT) [9] Zywno, M. ,”A Contribution to Validation of Score Meaning for Felder-Soloman's Index of Learning Styles”, ASEE Conference, Nashville, Tennessee,2003. [10] Korneliya Yordanova, “Meta-Data Application in Development, Exchange and Delivery of Digital Reusable Learning Content”, Interdisciplinary Journal of Knowledge and Learning Objects, Volume 3, p. 229-237, 2007. [11] C. Limongelli, F. Sciarrone, G. Vaste, “LS-Plan : An Effective Combination of Dynamic Courseware Generation and Learning Styles in Web-Based Education”, Proc. Fifth International Conference in Adaptive Hypermedia and Adaptive Web-Based Systems, p. 133-142, 2008. [12] C. Limongelli, F. Sciarrone, G. Vaste, “An Application of the LS-Plan System to an Educational Hypermedia”, Intl J Web-Based Learning and Teaching Technologies, vol.4. no.1, p. 15-34, 2009. [13] S Viola, S. R., Graf, S., Kinshuk, & Leo, T.,” Investigating relationships,within the Index of Learning Styles: A data-driven approach”, International Journal of Interactive Technology and Smart Education, Vol. 4, No.1, p. 7–18, 2007. [14] Huey-Ing Liu Min-Num Yang, “QoL guaranteed adaptation and personalization in E-learning systems”, Transactions on Education, Vol. 48, Issue4, p. 676 – 687, 2005. [15] Yi-Chun Chang et al., “A learning style classification mechanism for e-learning”, Computers & Education, Vol.53, p. 273-285,2009. [16] C.A. Carver, R.A. Howard, W.D. Lane, “Enhancing Student Learning Through Hypermedia Courseware and Incorporation of Student Learning Styles,” Transactions on Education, Vol.42, No.1, p. 33-38, February 1999. [17] E. Sanigneto, N. Capuano, M. Gaeta, A. Micarelli, “Adaptive Course Generation through Learning Styles Representation”, Universal Access Information Society(UAIS’08), Vol.7,No.1/2, p.1-23, 2008. [18] Karampiperis, P., & Sampson, D. (2005). Adaptive Learning Resources Sequencing in Educational Hypermedia Systems”, Educational Technology & Society, Vol. 8, No.4, p. 128-147. [19] Anitha D, Deisy C, “Deriving a prototype for the dynamic generation of learning path in an e-learning environment using Genetic algorithm”, International Journal of Innovation and Learning, Article in Press. Authors Anitha D is a research scholar of Anna University of Technology, Madurai, India and currently working as Assistant Professor in the Department of Computer Applications, Velammal College of Engineering & Technology,Madurai, India. She has done her Masters in Computer Applications and is interested in the application of Knowledge based techniques in the field of E-learning. She firmly believes that improving e-learning will enhance the teaching-learning process thereby improving the potential of excellence in educational institutions. Dr. C. Deisy received the Ph.D. degree in Computer Science from Anna University, Chennai in 2010. She is currently working as Associate Professor in the Department of Computer Science and Engineering, Thiagarajar College of Engineering ,Madurai, India. She has authored numerous scientific publications. Her main research interests are in the field of Data Mining and its applications. She has published her research works in many International Journals and conferences. She is the Journal Reviewer of Elsevier Journal of Applied Soft Computing since 2010.