An efficient hierarchical graph clustering algorithm based on shared neighbors and links

Abstract

Community structure is an important property of networks. A number of recent studies have focused on community detection algorithms. In this paper, we propose an efficient hierarchical graph clustering algorithm based on shared neighbors and links between clusters to detect communities. The basic idea is that vertices in the same cluster should have more shared neighbors than that in different clusters. We test our method by computer generated graphs and compare it with MCL algorithm. The performance of our algorithm is quite well. © 2013 Springer-Verlag Berlin Heidelberg.