Community Evolution in Dynamic Social Networks

Abstract

This paper proposed a framework and an algorithm for identifying communities in dynamic social networks. In order to handle the drawbacks of traditional approaches for social network analysis, we utilize the community similarities and infrequent change of community members combined with community structure optimization to develop a Group-based social community identification model to analyze the change of social interaction network with multiple time steps. According to this model ,we introduced a greed-cut algorithm and depth-search-first approach and combine them to develop a new algorithm for dynamic social interaction network recognition (called ADSIN). In addition, we conduct experiments on the dataset of Southern Women, the experiment results validate the accuracy and effectiveness of ADSIN.