A new Community Detection algorithm based on Distance Centrality

Community detection is important for many complex network applications. A major challenge lies in that the number of communities in a given social network is usually unknown. This paper presents a new community detection algorithm-Distance Centrality based Community Detection (DCCD). The proposed method is capable of detecting the community of network without a preset community number. The method has two components. First we choose the initial center nodes by calculating the centrality of each node using their distance information. Then we measure the similarity between the center nodes and each other nodes in the network, and assign each node to the most similar community. We demonstrate that the proposed distance centrality based community detection algorithm terminated on a good community number, and also has comparable detection accuracy with other existing approaches.