TagRec: Leveraging Tagging Wisdom for Recommendation

Due to the exponential growth of information on the Web, Recommender Systems have been developed to generate suggestions to help users overcome information overload and sift through huge amounts of information efficiently. Many existing approaches to recommender systems can neither handle very large datasets nor easily deal with users who have made very few ratings. Moreover, traditional recommender systems consider only the rating information, resulting in the loss of flexibility. Tagging has recently emerged as a popular way for users to annotate, organize and share resources on the Web. Several research tasks have shown that tags can represent userspsila judgments about Web contents quite accurately. In the light of the facts that both the rating activity and tagging activity can reflect userspsila opinions, this paper proposes a factor analysis approach called TagRec based on a unified probabilistic matrix factorization by utilizing both userspsila tagging information and rating information. The complexity analysis indicates that our approach can be applied to very large datasets. Furthermore, experimental results on MovieLens data set show that our method performs better than the state-of-the-art approaches.