Extracting Community Structure of Complex networks by Self-Organizing Maps

Abstract Identifying community structure is an important issue in network science and has attracted attention of researchers in many fields. It is relevant for social tasks, biological inquires, and technological problems. In this paper, we proposed a new approach based on self-organizing map to community detection. By using a proper weight-updating scheme, a network can be organized into dense subgraphs according to the topological connection of each node. Besides unweighted undirected networks, our method can also be used to detect communities in both weighted and bipartite networks.