Labeling Communities Using Structural Properties

Mobile Social Network Analysis is the mapping and measuring of interactions and flows between people, groups, and organizations based on the usage of their mobile communication services. Social Network Analysis and Mining has been highly influenced by the online social web sites, telecom consumer data and instant messaging systems, and has widely analyzed the presence of dense communities using graph theory and machine learning techniques. Community mining is one of the recent major directions in social network analysis. In this paper we find the communities in the network based on a modularity factor. Then we propose a graph theory based algorithm for further split of communities resulting in smaller sized and closely knit sub-units, to understand consumer behavior in a comprehensive manner. These sub-units are then analyzed and labeled based on their group behavior pattern. In this paper we measure and analyze the uniqueness of the structural properties for each small unit, it is another quick way to assign suitable labels for each distinct group. The effectiveness of the employed algorithms was evaluated on a huge telecom database in three different stages of our work.