文献类型：期刊文献

中文题名：基于SLPA优化的重叠社区发现算法

英文题名：AN IMPROVED OVERLAPPING COMMUNITY DETECTION ALGORITHM BASED ON SLPA

作者：陈界全[1];王占全[1];李真[2];汤敏伟[2]

机构：[1]华东理工大学信息科学与工程学院,上海200237;[2]天翼电子商务有限公司风险管理部,上海200080

年份：2021

卷号：38

期号：1

起止页码：297

中文期刊名：计算机应用与软件

外文期刊名：Computer Applications and Software

收录：CSTPCD;;北大核心:【北大核心2020】；

语种：中文

中文关键词：复杂网络;社区发现;重叠社区

外文关键词：Complex networks;Community detection;Overlapping community

摘要：传统的重叠社区发现算法SLPA虽然具有时间复杂度和性能上的优势,但标签传播算法内在的随机策略使得算法结果并不稳定。针对SLPA的缺点,提出一种高效稳定的重叠社区发现算法L-SLPA。先对网络进行非重叠划分,减少不同标签分配的数量,同时加入边界节点的考虑进行剪枝,以提高运行速度。实验结果表明,相比于SLPA,该算法在降低运行时间和随机性的同时保证了结果的准确性。
As a traditional overlapping community detection algorithm,SLPA has the advantages of time complexity and performance.However,its results are unstable due to the inherent random strategy of label propagation algorithm.For the purpose of addressing the shortcomings of SLPA,we propose an efficient and stable overlapping community detection algorithm L-SLPA.The network was preliminarily divided into non-overlapping communities to reduce the number of different label assignments.Meanwhile,we pruned the algorithm with the consideration of boundary nodes to improve the running speed of the algorithm.The experimental results show that compared with SLPA,L-SLPA reduces the running time and randomness while ensuring the accuracy of the results.