Modularity Based Community Detection in Dynamic Social Networks

Regular change is one of the features of network observed while analyzing dynamic social networks. But it is not easy to find community detection on real world dynamic social network. Dynamic social network generators are used insertion/deletion for node/edge and merge/split of community. In this paper, a method for community detection in a dynamic social network is proposed. Initially we take a static network, by moving a clique from one community to another community and flipping edge method in static network. By repeating this step, we get a dynamic network with improved modularity at every time step. So this method is used for community detection in dynamic network. An experiment is conducted to study changes in various network parameters like number of node, clustering coefficient, average distance between nodes etc.