A Novel and Integrated Semantic Recommendation System for E-Learning using Ontology
Mugdha Sharma, Laxmi Ahuja · 2016
As e-learning becomes increasingly popular among learners, the amount of course content on the Web also keeps expanding. Hence finding the most appropriate content for oneself has become a time-consuming task. Thus new systems which can efficiently recommend the most appropriate content to users based on their preferences are in demand. As a step towards providing the learners with such a system, we propose an ontology-based integrated recommendation system in the e-learning domain. Two ontologies viz. learner ontology and learning domain ontology are utilized to represent and model the knowledge about the learner and the learning domain respectively. In order to recommend the most appropriate content to users, our recommendation technique integrates four basic approaches: User profile matching, prerequisite learning path construction, collaborative filtering and semantic similarity technique. During their early phases, the recommender systems generally face the cold-start problem. This is because of the scarcity of information available during these phases. Our system successfully overcomes this problem by maintaining an ontological approach to user profiling. Our system also significantly improves the accuracy of recommendations, which we demonstrate experimentally.