Multiple Voting Label Propagation Algorithm for Overlapping Communities Detection

In order to fix the strong randomness and low-accuracy issue of traditional label propagation algorithm (LPA) and extend it into overlapping communities' detection, a new multiple voting label propagation algorithm is proposed. Firstly, potential leaders are chosen out based on node's degree and local covering density. And then, the successors are determined based on voting mechanism. At last, overlapping communities are detected with multiple voting labels. The simulation results show that this new method not only solves the defects of traditional LPA, but also has higher validity and feasibility.