An overlapping community detection algorithm with label propagation control for complex networks

Aiming at the problem that the accuracy of community detection is unstable and the labels appear vibration in the traditional overlapping community detection methods based on label propagation, this paper proposed OLPC (An Overlapping Community Detection Algorithm with Label Propagation Control for Complex Networks). The algorithm firstly initializes the labels and corresponding storage space for every node in networks. Then setting the number of reserved labels in the label storage space of nodes. And analyzing whether the node needs to continue the update operation in the way of judging whether the reserved labels in the storage space are same. Afterwards, every node receives the most appropriate community label by analyzing the neighbor nodes’ conditions. Finally, if the newest community labels received by all nodes are consistent with all nodes’ community labels received by previous generation’s label propagation, the algorithm stops. Through the testing in benchmark networks, real-world networks and the analysis after comparing the algorithm with some typical algorithms, the experimental results verified the feasibility and validity of the algorithm proposed in this paper. © 2019 Lavoisier. All rights reserved.