SocioRank*: A community and role detection method in social networks

Abstract

Due to the widespread popularity of cyber-social networks, research on community and social role discovery has drawn immense attention. Most of the available literature consider community and social role detection as areas of a disjoint problem domain. Moreover, existing research on concurrent community and social role detection utilizes a predefined set of roles and lacks in providing enough ground truth information for the proof. Hence, there is a high need for community structure analysis from an independent social perspective. Therefore, we propose a community and social role detection method called SocioRank* using email interaction data crawled from a graduate class of students. SocioRank* detects the underlying communities and ranks individuals on the basis of degree, closeness, and betweenness centrality. We perform extensive graph metrics-based experiments to evaluate the effectiveness of the proposed method. The comparison of results with the ground truth information demonstrates the high accuracy of SocioRank* method in community and role discovery.