Detecting overlapping and hierarchical communities in complex network using interaction-based edge clustering

Most community detection methods use network topology and edge density to identify optimal communities. However, in these methods, several objects that are connected by high weights may be decomposed into different communities, even when they intuitively belong to a single community. In this case, it is more effective to classify the objects into the same community because they perform important roles in controlling and understanding the network. To achieve this goal, in this paper, we propose a method of detecting optimal community structures in a complex network using interaction-based edge clustering. Our approach is to consider network topology as well as interaction density when identifying overlapping and hierarchical communities. Additionally, we measure the differences between the quantity and quality of intra- and inter-community interactions to evaluate the quality of the community structure. We test our method on several benchmark networks with known community structures. Additionally, after applying our method to several real-world complex networks, we evaluate our method through comparison with other methods. We find that the community quality and the overlap quality for our method surpass the results of the other methods.