CBACCN : Constraint Based Community Discovery in Complex Network

The prosperity of Web 2.0 and social media brings about many diverse social networks of unprecedented scales, which present new challenges for more effective graph-mining techniques. One of the most relevant features of networks representing real systems is community structure or clustering i.e. the organization of vertices joining with many edges lie in the same cluster and comparatively few edges joining vertices lie in different clusters. In this paper we also have used the notion of community as the key concept. In our previous work, we presented a divisive hierarchical clustering procedure based on minimum edge-cut principle and two constraints which have been used as stopping criteria in the partition phase of CHAMELEON. Here we improved that method by modifying our objective function to get more sophisticated community structures and follow centrality measure average clustering coefficient to measure quality of clusters. For some well-known benchmark complex networks, our method determines similar or more sophisticated communities with high average clustering coefficient than the existing methods of various research papers