TimeTrustSVD: A collaborative filtering model integrating time, trust and rating information

Collaborative filtering is widely used in recommender systems, which has better performance when trust relationship is taken into account as well as traditional rating information, as done in TrustSVD. However, the user preferences varies over time. Consequently, it is significant to take temporal information into consideration. This paper incorporates time influence both on users and items and proposed TimeTrustSVD, a collaborative filtering model integrating time information, trust relationships and rating scores. The proposed algorithm includes modified models of user bias and item bias based on a state-of-the-art algorithm TrustSVD, and achieved satisfactory results according to the experimental phenomena on a real-world dataset.