Recommender System: Collaborative Filtering of e-Learning Resources

The significant amount of information available on the web has led to difficulties for the learner to find useful information and relevant resources to carry out their training. The recommender systems have achieved significant success in the area of e-commerce, they still have difficulties in formulating relevant recommendations on e-learning resources because of the different characteristics of learners. Most of the existing recommendation techniques do not take these characteristics into account. This problem can be mitigated by including learner information in the referral process. Currently many recommendation techniques have cold start problems and classification problems. In this paper, we propose an ontology-based collaborative filtering recommendation system for recommending learners' online learning resources based on a decision algorithm (DA). In our approach, ontology is used to model and represent domain knowledge about the learner and learning resources. Our approach is divided into four parts: (a) the creation of an ontology for the representation of the learner's knowledge and learning resources (b) the calculation of the similarity of the assessments according to the ontology and the prediction for the learner concerned; (c) generating the K best items by the collaborative filtering recommendation engine and (d) applying the DA on the proposed items to generate the final recommendations for the targeted learner. [For the complete proceedings, see ED590269.]