Detecting Communities Based On Edge-Fitness And Node-Similarity In Social Networks

Network is a kind of effective method to present complex system. The social activities of human contain many data sets. Most of these data sets can be expressed as the network. With the development of technology, a wide variety of social applications appear in the human social life, which not only enrich the social life, and also provide important data for the study of social networks. Analysis of these data and their regularity is very meaningful. Thus, as one of the main research directions of social network analysis, community detection is an important subject.With the development of online social networks recently, many algorithms have been proposed that are able to assign each node to an appropriate community. The traditional approaches were always focused on the node community, while some recent studies have shown great advantage of method of edge community on some level. This paper presents a novel algorithm for discovering local communities in networks. By setting an original edge as a seed, a constantly maximize fitness function is used to obtain a local edge community. Meanwhile, the method can effectively control the scale and scope of the local community based on the boundary node identification, so by this way the complete structure information of the local community is obtained. The algorithm has been tested on both synthetic and real world networks, and it has been compared with other community detection algorithms. The experimental results show significant improvement on detecting community structure.