Community Detection in Complex Networks Using Minimum Spanning Tree and Modularity Maximization
Subject Areas : electrical and computer engineering
1 -
2 -
Keywords:
Abstract :
Modularity is one of the prominent features of complex networks that divides the structure of these networks into community groups. So far, many methods have been used to identify communities in complex networks, but some of these methods have local optimizations that affect the order of processing nodes and the final solution. In this paper, a new method for finding communities in complex networks using split and merge is proposed. In this method, minimum spanning tree is used as a tool to detect dissimilarity between nodes. In the partitioning process, the edges that show the most dissimilarity are removed in the minimum spanning tree to create smaller groups of nodes in a community. In the merging process, each group is merged with the neighboring group whose combination has the highest increase in modularity compared to other neighboring groups. The results of experiments conducted on real networks and artificial networks show that the method proposed in this article has a good accuracy for identifying communities in complex networks.
