An optimization of collaborative filtering personalized recommendation algorithm based on time context information

Abstract

This paper proposes an improved collaborative filtering algorithm based on time context information. Introducing the time information into the traditional collaborative filtering algorithm, the essay studies the changes of user preference in the time dimension. In this paper the time information includes three aspects: the time context information; the interest decays with the time; items similarity factor. This paper first uses Pearson correlation coefficient calculates time context similarity, pre-filtering the time-context. Through the experiment, the improved algorithm has higher accuracy than the traditional filter algorithms without time factor in the TOP-N recommendation list. It proves that time-context information of user's can affect the user's preference.