A community detection algorithm based on community size

Community detection can help us understand the structure and functions of complex networks from the view of modularity. Previous algorithms of community detection represent some defects, such as high time complexity, limitation of practical application and unsuitability for large networks. Thus an improved algorithm by CNM algorithm based on community size, CNMCS algorithm, is proposed in this article. According to the data of authoritative IP-level IPv6 networks from Jan.2009 to Dec.2010 provided by CAIDA, CNMCS algorithm is applied to these real-world networks and compared with the performance of previous algorithm. The comparison results indicate that CNMCS algorithm represents better performance according to the analysis of divided communities and modularity.