A Kind of Collaborative Filtering Algorithm Based on User Clustering and Time Stamp

In the case of user rating data is very large and sparse, the effect of traditional collaborative filtering algorithm in use recommendation is unsatisfactory. In this paper, a collaborative filtering algorithm based on user clustering and time stamp is proposed. Which first clusters the users according to user's preferences for different types of items, thus user will only recommended items to members belong to his/her own cluster, thus can largely reduce the input data of recommendation algorithm, and can improve recommendation effectiveness. Then taking account the factor that user will change his/her interest, in the traditional collaborative filtering algorithm, time factor is added to realize a real-time user recommendation. Experimental results show that the proposed algorithm can improve the accuracy and recall rate of user recommendation.