A Time-Enhanced Collaborative Filtering Approach

Abstract

Collaborative filtering can predict an active user's interests for unrated items based on his observed ratings, and the issue of concept drift exists in most of recommender systems. Aiming at the issue of concept drift, a time-enhanced collaborative filtering approach is proposed in this work, in which a time weight is introduced into the framework of collaborative filtering. As the experimental results show, the proposed approach improves the recommendation accuracy in contrast with the basic collaborative filtering.