A mining algorithm for overlapping community structure in networks

摘要

Nowadays, most existing community detecting algorithms treat smaller communities within a society network independently to reduce the complexity. Yet in most real world cases, those communities are always interleaved and overlapping. To address this problem, a novel algorithm based on greedy algorithm, namely MA-OCS, is proposed, which adopts agglomerative method to extract overlapped communities within a society network. Experiments on different standard datasets indicate that the proposed algorithm excels existing GN and KL algorithms in speed and accuracy.