A Graph based Learning Path Recommendation Model for Adaptive Personalized Learning Environments

Abstract

E-learning Recommender Systems are gaining great importance nowadays due to newlinetheir ability to enhance the learning experience by providing tailor-made services based on newlinelearner preferences. The main focus of a Personalized Learning Environment is to newlineunderstand and adapt to the learners needs. Learners have different individual needs, goals, newlineand preferences that affect their learning process. Similarly, different learners have different newlinecharacteristics regarding learner s background knowledge, learners history, competency newlinelevel, learning style and learning activities. This difference in learner characteristics makes newlinethe recommendation of learning resources to a particular learner more difficult. newlineOne solution to this problem is integrating knowledge about the learner and learning newlineresources in the recommendation process. This thesis proposed an adaptive learning path newlinerecommendation method, which suggests a learning path, an ordered set of cognitively newlineconnected learning materials according to the learning need. In this research, ontology is newlineproposed to store the knowledge about learners and learning resources due to its dynamic newlinenature and knowledge-sharing capability across the domain. The thesis presents different newlinelearning material recommending methods for achieving dynamicity and adaptivity in newlinepersonalized content recommendations. The experimentation is done in an incremental newlinedevelopment technique which forms a serial analysis of system from less complex to more newlinecomplex procedures. A simple and primitive rule-based static recommender is implemented newlinein the first phase of the work. Using these rules in the next phase, an ontology framework is newlinedeveloped to store the student model and learning object features. The ontology is queried newlineto retrieve the similar learner groups based on their historical data. A sequential pattern newlinemining algorithm is implemented over the ontology-based approach to check the newlineeffectiveness of recommending learning sequences. newline