Community Detecting Algorithm For Complex Networks

Nowadays, people have been living in a world filled with all sorts of complexnetworks. The community structure of the complex networks is one of the mostcommon and important topological properties. In this structure, the same cluster nodesconnect to each other closely, while different cluster nodes interconnect sparsely. Thecommunity discovery method which is expected to reveal the the community structureof complex network is of important theoretical significance in analysis of topologystructure in complex network and is helpful in finding hidden patterns and predictionsof its behavior. It has been widely applied in social networks, biological networks andthe World Wide Web.In this paper, we do deep research and analysis on a variety of classic communitydiscovery algorithms in complex networks. In order to address the limitations of theclassical algorithms, a new definition and a new evaluation standard of the complexnetwork community structure are given. Weather a community is close depends on thelinks of nodes in it. The quality of community structure is measured by the differencesizes of edges in communities and edges between communities.Based on the new definitions, a new community detecting algorithm is proposed.The algorithm gets an initial community by finding a super dense core area in thenetwork, and then adds nodes to the community that is closely linked to it.finally thealgorithm merges and adjusts communities.necessarily. The algorithm not only does notneed numbers of communities and termination conditions set by man, but also can dealwith different scale datasets effectively.At last, its reasonability and efficiency havebeen verified through simulation experiments and the results show that the proposedalgorithm is effective.