Exploiting relational tag expansion for dynamic user profile in a tag-aware ranking recommender system

Abstract

A tag-aware recommender system (TRS) presents the challenge of tag sparsity in a user profile. Previous work focuses on expanding similar tags and does not link the tags with corresponding resources, therefore leading to a static user profile in the recommendation. In this article, we have proposed a new social tag expansion model (STEM) to generate a dynamic user profile to improve the recommendation performance. Instead of simply including most relevant tags, the new model focuses on the completeness of a user profile through expanding tags by exploiting their relations and includes a sufficient set of tags to alleviate the tag sparsity problem. The novel STEM-based TRS contains three operations: (1) Tag cloud generation discovers potentially relevant tags in an application domain; (2) Tag expansion finds a sufficient set of tags upon original tags; and (3) User profile refactoring builds a dynamic user profile and determines the weights of the extended tags in the profile. We analysed the STEM property in terms of recommendation accuracy and demonstrated its performance through extensive experiments over multiple datasets. The analysis and experimental results showed that the new STEM technique was able to correctly find a sufficient set of tags and to improve the recommendation accuracy by solving the tag sparsity problem. At this point, this technique has consistently outperformed state-of-art tag-aware recommendation methods in these extensive experiments.