Community detection in social tagging systems based semantics of tags

Community detection is an important technique in social tagging systems (STSs), which is of great value to personalized recommendation, information retrieval and user behavior analysis. In this paper, in terms of the agglomerative hierarchical clustering (AHC), a community detection algorithm based on semantics of tags is proposed, named as I-ACST (an improved agglomerative clustering based on semantics of tags). First, we directly use t ag co-occurrence to measure the tag semantic relevance, and then tags are merged using I-ACST algorithm. Finally, the proposed algorithm is tested on real network, the experimental results show that I-ACST algorithm enhances the clustering performance than other clustering algorithms, and verifies the availability and effectiveness.