A two-stage hierarchical community structure detected algorithm in complex network

Series of algorithms have been devised in detecting community. To make the detected community structure match better with the real network, we propose a new hierarchical community detection algorithm called Perm-Mod. The algorithm includes two stages. In the first stage, using the heuristic method, we optimize an objective function. When the function achieves maximum value, a few small communities are precisely detected. In the second stage, we estimate whether the network has community structure based on the maximum value. If yes, we merge the above small communities. Compared to other algorithms, our algorithm solves three problems at the same time: (i) we estimate whether the network has community structure (ii) the community structure found by our algorithm matches well with the ground-truth network and (iii) the process of merging reflects the hierarchy.