Toward a fully automatic learner model based on web usage mining with respect to educational preferences and learning styles Mohamed JEMNI Olfa NASRAOUI Mohamed Koutheaïr KHRIBI University of Tunis University of Louisville University of Tunis mohamed.jemni@fst.rnu.tn olfa.nasraoui@louisville.edu mk.khribi@uvt.rnu.tn Kinshuk Sabine GRAF Athabasca University Athabasca University kinshuk@athabascau.ca sabineg@athabascau.ca LaTICE, University of Tunis ICALT 2013, Beijing, China.
Learner Modelling Approaches Collaborative Automatic - Advantages of automatic learner modeling : No additional work for learners ; Uses information from a time span; higher tolerance Allows dynamic updating of information 2 ICALT 2013, Beijing, China 15/07/2013
General Aim Building an automatic recommendation system for Learning Management Systems. (Learning Objects) (Learners) Learner- Object LO LO 1 LO Interest Measures (Past& Recent) LO LO 2 LO i LO Usage Data Implicit LO n LO Recommendation Content Modeling (content, collaborative, hybrid) 2 1 3 Learner Features LO Features Learner Modeling 3 ICALT 2013, Beijing, China 15/07/2013
Research Question How to automatically model learners and groups of learners based on implicit data from their interactions and online activities in the learning management system, taking into account the learners’ * educational preferences * learning styles 4 ICALT 2013, Beijing, China 15/07/2013
General Features of the Proposed Approach Automatic, dynamic and based solely on learner usage sessions ; Input data is composed from collected implicit data tracked and saved in LMS database and/or web server log files ; Based on web usage mining techniques ; Educational preferences and learning styles are considered and identified automatically. 5 ICALT 2013, Beijing, China 15/07/2013
Proposed Learner Model Proposed Learner model components : LM i = {PR i , LK i , LEP i } General student information such as Identification PR i : i th Learner Profile data, Demographic information, etc. Sequences of weighted visited learning objects i.e. LK i : i th Learner’s Knowledge vectors of visited learning objects or curriculum Component elements in which the student was interested (learner’s knowledge) LEP i : i th Learner’s Educational Learner’s educational preferences and Learning Component style. 6 ICALT 2013, Beijing, China 15/07/2013
Learner’s Knowledge Component 7 ICALT 2013, Beijing, China 15/07/2013
Learner’s Knowledge Component 8 ICALT 2013, Beijing, China 15/07/2013
Learner’s Knowledge Component The learner’s knowledge component LK i can be represented as a matrix M(p, n), where p is the total number of learner’s sessions and n the cardinality of unique visited learning objects. LO 1 LO 2 LO 3 LO 4 … LO n LK i = 9 ICALT 2013, Beijing, China 15/07/2013
Learner’s Educational Component Composed of the learner’s preferences among visited learning objects and his/her learning style . Detection of the learner’s preferences : What kind of learning object does a learner prefer? Learning objects available in LMS are characterized by many attributes (e.g. format), each of which may have several values (e.g. for format: text, image, video, etc.) that could be preferred or not by the learner. The preferences of a learner i upon these values can be represented, as a vector of interest measures : Attributes Related values Type_LO(Learning object type) {Resource, Activity} {Exercise, Simulation, Questionnaire, Shape_LO(Learning object Assessment, Forum, Chat, Wiki, Assignment} shape, if Type_LO = Activity) Format_LO(Learning object {Text, HTML, Image, Sound, Video} format, if Format_LO = Resource) 10 ICALT 2013, Beijing, China 15/07/2013
Learner’s Educational Component 11 ICALT 2013, Beijing, China 15/07/2013
Learner’s Educational Component Attributes and corresponding Interest measures values Resource LOIM_Type_LO i Resource Type_LO LOIM_Type_LO i Activity Activity LOIM_Shape_LO i Exercice Shape_LO Exercice LOIM_Shape_LO i Questionnaire Questionnaire test LOIM_Shape_LO i Test LOIM_Shape_LO i Forum Forum LOIM_Shape_LO i Chat Chat Wiki LOIM_Shape_LO i Wiki LOIM_Shape_LO i Navigation Navigation LOIM_Shape_LO i example Exeamle Text LOIM_Format_LO i Text Format_LO LOIM_Format_LO i Html Html LOIM_Format_LO i Image Image Audio LOIM_Format_LO i Audio LOIM_Format_LO i Vidéo Video 12 ICALT 2013, Beijing, China 15/07/2013
Learner’s Educational Component Detection of the learning style (Graf et al., 2008) Sensing/Intuitive Commonly incorporated features in LMS Actifve/Reflective • Content • Outline • Example • Self-Assessment F S L S M • Exercice • Forum • Navigation Visual/Verbal Sequential/Global Active/Reflective Sensing/Intuitive Visual/Verbal Sequential/Global Content_visit(-) Example_visit(+) Ques_text(-) Outline_stay(-) Content_stay(-) Example_stay(+) Forum_visit(-) Ques_detail(+) Outline_stay(-) Selfass_visit(+) Forum_stay(-) Ques_overview(-) Selfass_stay(-) Selfass_stay(+) Forum_post(-) Ques_interpret(-) …. …. …. …. …. …. …. …. 13 ICALT 2013, Beijing, China 15/07/2013
Group Modeling LP i , {(LEP 1 , .. , LEP p ), LS i } , Once learner models are built, we apply a hierarchical multi- level model based collaborative filtering approach on these models, in order to assign learners with common preferences and interests to the same groups , so that feedback from one learner can serve as a guideline for information delivery to the other learners within the same group. 14 ICALT 2013, Beijing, China 15/07/2013
Level1 : Classification LS 1 LS 2 LS 8 Learning Styles … Educational Preferences Level2 : Clustering … … … … W1 W2 W3 . . Group vectors Level3 : Clustering . Wi .. Wm … … … … 15 ICALT 2013, Beijing, China 15/07/2013
Implementation and Proof of Concept Evaluation The proposed approach has been implemented and an experiment was performed as part of a recommendation approach Recommendations are computed with respect to the learner’s clickstreams, his/her learning style and educational preferences, as well as exploiting similarities and dissimilarities among the learner models and educational content. Moodle LMS is used to implement the proposed approach. We used an online hybrid course (C2i) with 704 learners . Tracked data was successfully extracted from various Moodle tables (primarily from mdl_log table). 16 ICALT 2013, Beijing, China 15/07/2013
HARSYPEL Architecture Logs + DB Learners’ Models Builder Content Model Builder Usage data component 2 Cp2.1 LK Builder Cp3.1 LO converter Learning Styles Models component 3 Cp3.2 Terms Extractor Cp2.2 LEP Builder Cp3.3 Indexer Cp2.3 Groups Builder LO (Sessionizing) (Indexing | Retrieving) Implicit Query Extractor H A R S Y P E L Cp1.1 LO Vector Builder Fenêtre glissante ( SW LO ) Recommendation Engine Cp1.2 Terms’ Vector Builder Fenêtre glissante ( K termes pertinents) Cp4.1 Collaborative Recommender component 1 Cp1.3 LEP Extractor Extraction des préférences Configuration pédagogiques HARSYPEL Cp4.2 Content Recommender Module component 4 component 5 Cp4.3 Hybridizer Input Learner Output (Active Session) 17 ICALT 2013, Beijing, China 15/07/2013
Conclusions and Future Work Developed an approach for automatically modeling learners (groups) within LMS taking into account their educational preferences and learning styles Proposed approach falls within the scope of building an educational automatic hybrid recommender system providing suitable recommendations to learners for personalized technology enhanced learning ; Future work: Evaluation of the recommender system 18 ICALT 2013, Beijing, China 15/07/2013
Recommend
More recommend
