Ontology-based collaborative filtering recommendation algorithm

Abstract

E-learning system for knowledge points recommended primarily uses traditional collaborative filtering algorithm. Similarity calculation of knowledge points is often based on user rating above the intersection of knowledge points. The different semantic relations between knowledge points are not well considered, which results in the low recommended accuracy. This paper proposed an Ontology-based collaborative filtering recommendation algorithm, which could help users find the nearest neighbors even if the scores of knowledge points are little or zero. Through experiment, this algorithm was compared to traditional collaborative filtering recommendation algorithms. The new method achieved a better recommendation. © 2013 Springer-Verlag.