关键词: 流式图, 重叠社区发现, 边聚类系数, 边图, 社区相似度

Abstract: Overlapping community detection is a major topic in complex network research. To improve the time efficiency of detection overlapping communities in the network, it proposes a link-based streaming overlapping community detection algorithm called LBSA. The algorithm firstly deals with the edges of the link graph in random order to get the initial communities, then merges the small overlapping communities with other the most similarly large communities. Through those two steps, the algorithm can get the overlapping structure of network with nearly linear time complexity. Experimental results show that the algorithm can detect the overlapping communities in the network with better time efficiency and has good performance compared to the other algorithms.

Key words: graph streaming, overlapping community detection, edge clustering coefficient, link graph, community similarity