Overlapping community detection in complex networks based on the boundary information of disjoint community

We propose an algorithm for overlapping detection in complex networks based on the boundary information of disjoint community. For a boundary node, it connects several disjoint communities. The algorithm computes the sharpness variations of their boundaries caused by the boundary node. It judges whether the boundary node belongs to multi-communities synchronously. Then the overlapping nodes among the disjoint communities can be found. Moreover, the overlapping ratio of complex networks can be controlled by parameter r of the algorithm. This is useful to reveal the overlapping communities of complex networks hierarchically to some extent. In experimental tests using an artificial random network and three real networks, the algorithm is proved to be simple and feasible, and it can detect overlapping communities in complex networks flexibly and effectively.