Centrality Approach for Community Detection in Large Scale Network

Identifying communities in social network plays an important role in predicting behavior of the complex network. Real world systems in social network can be modeled as a graph structure, where nodes represents the social entities and edges represents the relationships among the entities. Usually nodes inside a community are having similar kinds of properties and most of them are influence by one or more central nodes in the network. Hence centrality principle can be adapted for efficiently discovery of communities. In this paper, an attempt has been made for community detection using central nodes of the network. Discovering central nodes in large scale network is a challenging task due to its huge complex structure. Central nodes have been been identified using map reduce paradigm in order to carry out the computation in distributed manner. The process of discovering communities is then carried out using the identified central nodes. Experimental evaluation shows that the proposed method for community detection provides better performance in term of both accuracy and time complexity.