Comparison of fuzzy co-clustering methods in collaborative filtering-based recommender system

Abstract

Various fuzzy co-clustering methods have been proposed for collaborative filtering; however, it is not clear which method is best in terms of accuracy. This paper proposes a recommender system that utilizes fuzzy co-clustering-based collaborative filtering and also evaluates four fuzzy co-clustering methods. The proposed system recommends optimal items to users using large-scale rating datasets. The results of numerical experiments conducted using one artificial dataset and two real datasets indicate that, the proposed method combined with a particular fuzzy co-clustering method is more accurate than conventional methods.