Optimizing Community Detection in Complex Networks by Regrouping Gateways

Community detection that can extract potential hidden structures and key information of large-scale complex networks has been studied so many years. However, there exists a possibility of grouping the gateway incorrectly which is the central node communicates with the outside, thus resulting in a large loss of modularity. Here we propose a regrouping gateways (RG) algorithm for community detection to optimize the existing community partitions which have been divided by various algorithms. We design a heuristic optimization algorithm based on a breadth-first search strategy with less complexity than Leiden to improve modularity. Experiments carried out on multiple real networks show that RG exhibits good generality in effectively increasing the modularity and reducing the randomness of community detection algorithms.