Detecting Communities in Dynamic Social Networks using Modularity Ensembles SOM

Social network analysis has gained much importance these days. Social network analysis is the process of recording various patterns of interactions between a set of social entities. An important phenomenon that draws the attention of analysis is the emergence of communities in these networks. The understanding and detection of communities in these networks is a challenging research problem. However, approaches to detect communities have largely focused on identifying communities in static social networks. But real-world social networks are not always static. In fact, many social networks in reality (such as Facebook, Bebo and Twitter) are dynamic networks that frequently change over time. In this paper, a framework is proposed for community detection in dynamic social networks, which explores self-organizing maps (SOM) for cluster selection and modularity measure for community strength identification. Experimental results on synthetic network datasets show the effectiveness of the proposed approach.