A Personalized Recommendation Algorithm Based on the Fusion of Trust Relation and Time Series

In order to further improve the accuracy of personalized recommendation algorithm in social network, on the basis of summarizing the traditional recommendation algorithm, this paper introduces the social relationship between users, the trust propagation mechanism and time sequence information and user-item score matrix information are fused to the probability matrix decomposition model, a new personalized recommendation model TTSMF is established, the model learns the potential features of the user and the item, and consider the time factor, and handle trust relationship between users. Even if the user does not score on any item, it can also learn the user's feature vector by trusting relationship. Compared with existing algorithms, TTSMF algorithm can better solve the cold start problem and improve the accuracy of the algorithm. By analyzing the time complexity of the algorithm, the TTSMF algorithm can be easily extended to the application scenarios with large data sets.