Research on Personalized Recommendation Algorithm Based on Time Weighted and Sparse Space Clustering

Abstract

Collaborative filtering algorithm is one of the most successful recommendation algorithms in personalized recommendation system, but the traditional algorithm does not consider the user's interest changes in different time periods, resulting in the set of neighbors may not be the nearest neighbor set. what is more.because of data sparsity and computational complexity, the efficiency of the algorithm is poor. time-weighting and clustering appear as a nature solution to this problem. So this paper proposes a collaborative filtering algorithm based on users' interest in different time period. First.the algorithm performs sparse subspace clustering on users solving data sparsity problem and improving the accuracy for searching similar neighbors, then we assigns each item a score that gradually decreases with time using the weighted score and find the nearest neighbor of the target user. According to the history of similar friends' watching videos, recommend high rated videos to target users. We conduct experiments using real dates from movielens to verify our algorithm and evaluate its performance. Experiments demonstrate that the improved algorithm improves the recommendation quality of the collaborative filtering recommendation system.