Temporal Social Tagging Based Collaborative Filtering Recommender for Digital Library

Social recommendation is one of the exciting personalized services in digital libraries, which still faces the problems of cold start and dynamic interest transferring in traditional collaborative filtering algorithms. This paper proposes a hierarchical collaborative filtering recommendation algorithm based on the social tagging and the temporal interesting modeling. Firstly, the reduced user-book-tag tensor model is adjusted by the interest transferring curves, which fitted by the temporal tagging behavior of each tags. Then, the candidate social tags are extracted from the social community by the rebuild User-Tag matrix C. After constructing the user model and the item model by matrix factorization, the books with the highest posterior of the tags are recommended by the naive Bayes classifier. Experimental results show that the proposed algorithm improves the recommendation performance especially for the highly time-sensitive data.