Detecting Overlapping Community in Social Networks Based on Fuzzy Membership Degree

摘要

Overlapping community detection in social networks is a challenging task for revealing the community structure, as one user may belong to several communities. Most previous methods of overlapping community detection ignore the belonging levels when one node belongs to several communities. The membership-degree is used to embody the belonging level. In this paper, an novel method calling Fuzzy Membership-Degree Algorithm (FMA) is put forward. Firstly, we propagate the membership-degree with consideration of the nodes-attraction, which is a new proposed definition based on topological characteristics. Then we further mine communities under the guidance of Extended Modularity (EQ). In this paper, the proposed algorithm FMA makes full use of the topological information, and membership-degree suggests the belonging level of overlapping community. Experiments on synthetic and real-world networks demonstrate that our algorithm performs significantly.