An optimized overlapping community detection algorithm KC-SLPA is proposed based on the idea of la-bel propagation for overlapping community discovery on complex networks.By comprehensively considering the lo-cal and global attribute indicators of network structure,the node importance is measured by fusing the node influ-ence score of K-shell and clustering coefficient,reducing the randomness of the algorithm through the fixed access order of node influence ranking;an improved Speaker-Listener rule is proposed,which propagates labels based on the average frequency of label occurrences,avoiding the randomness problem of label selection;an improved node similarity is introduced to further process multiple Speaker labels,improving the quality of community detection.Experiments are conducted on both synthetic networks and real datasets,and the results show that the algorithm has high stability in networks of different sizes and detects overlapping communities with good quality.
complex networkoverlapping community detectionlabel disseminationK-shell
维普
万方数据
