An overlapping community detection algorithm based on density peaks

摘要

Many real-world networks contain overlapping communities like protein-protein networks and social networks. Overlapping community detection plays an important role in studying hidden structure of those networks. In this paper, we propose a novel overlapping community detection algorithm based on density peaks (OCDDP). OCDDP utilizes a similarity based method to set distances among nodes, a three-step process to select cores of communities and membership vectors to represent belongings of nodes. Experiments on synthetic networks and social networks prove that OCDDP is an effective and stable overlapping community detection algorithm. Compared with the top existing methods, it tends to perform better on those "simple" structure networks rather than those infrequently "complicated" ones.