Overlapping Community Detection Algorithm Based on Edge Strength

Communities represent an ubiquitous topological characteristic of complex networks, and discovering community structures is of fundamental importance. Conductance is a detection algorithm for weighted overlapping communities with high-accuracy division results; however, the relationship between the nodes and their neighbors is not considered in the selection of the initial community, which leads to unreasonable initial community selection and lower accuracy in discovering the real community structure of the network. In addition, the algorithm may miss nodes. Accordingly, the edge strength conductance algorithm (ESCA) is proposed, which resolves the issues of unreasonable initial community selection and missing nodes by using the concepts of edge strength and belonging degree. Experiments demonstrate that for both unweighted and weighted networks, ESCA does not miss nodes, and the detected communities are closer to the real network community structure compared with those obtained by Conductance and COPRA, the community overlap propagation algorithm.