A game-theoretic approach for non-overlapping communities detection

In this paper, we propose a game-theoretic approach to find the community structure in complex networks based on a non-cooperative game. This approach optimizes a node-based modularity for non-overlapping communities. Experiments show that our approach is effective to discover non-overlapping communities and obtain high values of modularity and Normalized Mutual Information (NMI) for real-world and synthetic networks in a reasonable time. © 2017 IEEE.