Overlapping Community Detection by Local Community Gravitation in Social Networks

In this paper, a simple but efficient method of overlapping community detection is presented using local community gravitation in social networks. Given a highquality, non-overlapping partition generated by existing methods, this proposed method identifies the overlapping nodes from their surrounding partitioned communities according to their local community gravitation, with low computational complexity. Our experiments on synthetic networks and real-world networks demonstrate that the proposed algorithm is better than other algorithms in terms of the general quality.