Influence spreading model used to community detection in social networks

Abstract

A new method for community discovery in a social network is presented. The proposed algorithm computes local maxima of an influence measure where influence of nodes is considered in both directions. A node has influence on others and a node is influenced by others. Network topology, and possible directed connections and unequal weights of nodes and links, are essential features of the model. Modeling can be adjusted by model parameters giving flexibility in various applications. Weaker connections give rise to more sub-communities where as stronger ties increase the cohesion of the community. The validity of the method is demonstrated by two different social networks.