A Graph Theory Based Method to Extract Social Structure in the Society

Abstract

Huge amount of data is being produced by online social interaction among people. This data can be represented by graphs where nodes represent individuals and the connecting edges depicts their interaction. In this research we analyze a social network of individuals to understand social structure among them. Online interaction has become integral part of life nowadays, large amount of research is available in analyzing these social interactions. However, previous research in this area lacks in identifying social structure among people using email data and graph theory based techniques. In this regard, a model for analyzing social structure of the community is presented in this research. An algorithm is designed to extract social structure of the community named Socio Rank. We crawled a large real world email interaction data in this research and extensive graph theory based experiments are performed to cluster the graph among different communities. Subsequently, widespread analysis was performed to study the hierarchy of social structure in the society. The experiments revealed multiple clusters in the group related to individuals fulfilling different roles in the community. We correlated the connection properties of individual nodes with the behavior of people in the society. Graph based Harel-Koren layout technique and Girvan-Newman clustering algorithm was used for the analysis of the underlying extracted communities. A hierarchical social classification was identified among individuals in the community. Our work of social structure extraction on a controlled community can be correlated with the society at large.