Abstract: The local expansion-based overlapping community detection algorithms start from a single node or a clique,and then repeatedly add a new node to the obtained community,thus obtaining the final community.However,the existing local expansion-based overlapping community detection algorithms always add one node at each step,so the local information of the added nodes has not been fully considered,thus reducing the accuracy of the algorithms.In order to solve this problem,an overlapping community detection algorithm is proposed,which adds a clique at each step.This algorithm starts from a clique,and expands the clique by repeatedly adding a clique of the largest fitness value in the neighborhood.In this way,the links between the added node and the existing community and the links between the added nodes are both taken into account.An empirical evaluation on the synthetic and real datasets demonstrates that the proposed algorithm can detect the overlapping communities in complex networks more accurately than the existing algorithms.

Key words: complex networks, community detection, overlapping community, local expansion