A Fast Algorithm for Finding Community Structure Based on Community Closeness

Recently, the characterization of community structures in complex networks has received a considerable amount of attentions. Effective identification of these communities or clusters is a general problem in the field of data mining. In this paper we present a fast hierarchical agglomerative algorithm based on community closeness (FHACC) algorithm, for detecting community structure which is very efficient and faster than many other competing algorithms. FHACC tends to agglomerate such communities that share the most common vertices into larger ones. Its running time on a sparse network with n vertices and m edges is O(mk+mt), where k denotes the mean vertex degree, and t is the iteration times of community agglomeration in FHACC algorithm. The algorithm was tested on several real-world networks and proved to be high efficient and effective in community finding.