An overlapping community detection algorithm based on triangle coarsening and dynamic distance

Abstract

Discoverying hidden communities in various kinds of complicated networks is a considerable research direction in the field of complex network analysis. Its goal is to discover the structures of communities in complex networks. The algorithms devised upon the Attractor dynamic distance mechanism are capable of finding stable communities with various sizes. However, they still have deficiencies in overlapping community discovery and runtime efficiency. An overlapping community discovery algorithm based on triangle coarsening and dynamic distance is posed in this paper. First, a coarsening strategy devised upon triangle is adopted to reduce networks’ sizes. Second, for the coarsened networks, a dynamic distance processing mechanism based on overlapping Attractors is used to discover the overlapping communities in the networks. Finally, the communities in the raw networks are obtained through anti-roughening steps. The experiments on different datasets demonstrate that the proposed algorithm not only can discover the overlapping communities accurately but also has low time complexity.