A New Similarity Measure-Based Collaborative Filtering Approach for Recommender Systems

Collaborative filtering (CF) is the most popular recommendation approach in personalization techniques but still suffers from poor recommendation accuracy. This study incorporates fuzzy set technique and user-relevant analysis to improve the CF approach. It proposes an innovative fuzzy similarity measure (FSM) and user-relevant aggregation (URA) on recommendation approach. Experiments demonstrate that the FSM-URA approach significantly improves the prediction accuracy comparing to the existing recommendation approaches. © Springer-Verlag Berlin Heidelberg 2014