Similarity Aware Community Detecting Method for Social Intermittently Connected Mobile Network

With the store-carry-forward transmitting manner in social intermittently connected mobile network, the nodes always behave in community movement model. A community structure detecting method is proposed based on similarity and community label propagation. According to the statistics of roaming epochs in community movement, the centrality of nodes can be evaluated accurately; moreover, the network update correlative situation after controlling the label propagation with certain constraint condition. Finally the whole network is divided into several communities logically, and centrality nodes are also probed. Simulation results show that the method is effective.