Finding Community Structure in Social Network by Electric Circuit Modeling

The community structure is an important property of social networks. Motivated by circuit theory, a social network can be represented as an electric circuit with notions of voltages of nodes and currents of edges. Based on the electric circuit model, we propose an improved algorithm for finding the community structure in social networks. In the proposed algorithm, a social network can be divided into communities with modularity maximization by repeatedly removing the edge that has the largest current in the corresponding electric circuit after random voltage initializations. The proposed algorithm is tested against three benchmark datasets of real social networks with either unweighted or weighted network graphs. The tests demonstrate the advantages of the proposed algorithm including high accuracy, low complexity, and no prior information needed.