Discovering generalized communities in weighted networks

Recent years have witnessed the rapid development of community detection and a large collection of algorithms has been proposed. As weights may carry crucial information, community detection in weighted networks has also attracted the researchers' interest. In this paper, an algorithm is proposed to discover generalized communities of more structural patterns in weighted networks, including overlapping communities, disassortative structure, etc. It considers the network weights in the modeling and computation to study the inference of the latent continuous structures in weighted networks. The algorithm is tested both on the benchmark graphs and the real-world network. Results show good performances and favorable properties of the algorithm. Copyright (C) EPLA, 2016