A New Community Division based on Coring Graph Clustering

Abstract

A new community finding algorithm, based on the greedy algorithm with graph clustering by computing the density variation sequence and identifying core nodes, number of communities, partition the certain nodes to some belonged community with the similarity of characteristics of communication behavior by continuous readjusting the centrality of the communities. The use of community density and effective diameter to measure the quality of the community partition on the real datasets of email corpus shows the feasibility and effectiveness of the proposed algorithm.