Social Recommendation Algorithm Based on the Context of Time and Tags

Recommender systems are traditionally based on a 2-dimensional user-item ratings matrix and focus on recommending the most relevant items to users regardless of context. Therefore, the accuracy and reliability of conventional recommendation algorithms are relatively low. Social tags can realize the classification of information and are marked by users freely which can be used to mining users´ preference for items. What´s more, it´s easy to obtain time. In this paper, in order to enrich the individual user or item information, we take the time and tag context into consideration for the recommendation algorithm. The time context is used to mine the influence of relationship between the two users, and the tag context is used to measure the relationship between items at the same time. Finally, user relation vector and item relation vector are fused to the probabilistic matrix factorization model. Experimental results show that the proposed social recommendation algorithm in this paper can improve the accuracy and precision of recommendation.