A new method for detecting communities in network based on the affinity between nodes

Abstract

It has been found that many networks display community structure, analysis of the com-munity structure in the network is meaningful for understanding the real-world network. Recently, many community discovery methods have been proposed, most of which are based on the assumption that nodes in the network can only belong to a community at most. The proposed method in the paper is based on the affinity between nodes and it does not agree with the assumption, i.e., the method is able to discover overlapping community structures in networks. We apply the method to various real-world networks and artificial networks, the experimental results confirm that it has better performance on networks with overlapping community structures and non-overlapping community structures.