Exploring Information Hidden in Tags: A Subject-Based Item Recommendation Approach

Collaborative tagging sites allow users to bookmark and annotate their favorite Web contents with tags. These tags provide a novel source of information for collaborative filtering (CF). Research on how to improve item recommendation quality leveraging tags is emerging yet information hidden in tags is far from being fully exploited. In this paper, we aim at finding informative usage patterns from tags by consistent clustering on tags using nonnegative matrix factorization. The clustered subjects, represented by weighed tag vectors, can then be used to build a subject-centered user information seeking model for item recommendation. Experiments on two real-world datasets show that our subject-based algorithms substantially outperform the traditional CF methods as well as tag-enhanced recommendation approaches reported in the literature.