An Overlapping Community Detection Algorithm Based on Link Clustering in Complex Networks

Community detection has important significance for understanding network topology and analyzing network function. It has been shown that there are high overlapping community structures in the complex networks. However, it is difficult to detect these structures for the existing community detection algorithms. This paper proposes an algorithm (CLCD) to detect high overlapping community structures. This algorithm starts from the perspective of the link. Through selecting a core link, this algorithm attracts links in the outer space to join in the community which contains the core link only in the beginning. Finally, transform link communities to node communities. The global optimal overlapping community structures will be formed after adjusting the node communities. This algorithm can get the number of communities automatically without inputting additional parameters. The examples of application to both artificial networks and real networks give better results on detecting high overlapping community structures.