融合节点影响力与多标签传播的重叠社区发现算法  

Overlapping community discovery algorithm integrating node influence and multi-label propagation

作　　者：陈荣旺 江彩英 郭昆[3] CHEN Rongwang;JIANG Caiying;GUO Kun(College of Mathematics and Computer Science,Wuyi University,Nanping,Fujian 354300,China;Information Network and Equipment Support Center,Nanping Meteorological Bureau,Nanping,Fujian 353000,China;College of Computer and Big Data,Fuzhou University,Fuzhou,Fujian 350108,China)

机构地区：[1]武夷学院数学与计算机学院,福建南平354300 [2]南平市气象局信息网络与装备保障中心,福建南平353000 [3]福州大学计算机与大数据学院,福建福州350108

出　　处：《福州大学学报（自然科学版）》2023年第4期451-458,共8页Journal of Fuzhou University(Natural Science Edition)

基　　金：国家自然科学基金区域联合重点项目(U21A20472);福建省自然科学基金资助项目(2020J01420)。

摘　　要：针对传统多标签传播算法存在的社区发现结果具有随机性、不稳定性,以及忽视节点影响力对标签传播的影响等问题,提出一种基于节点影响力与多标签传播的能够生成稳定社区的重叠社区发现算法.算法在节点影响力的计算、排序和核心节点识别基础上,通过邻居节点初始标签的再处理和基于平衡系数的节点标签异步更新策略,实现复杂网络重叠社区的有效识别.在真实数据集和人工数据集上的实验综合表明,算法性能优于各对比算法,适用于大规模复杂网络.Traditional multi-label propagation algorithms suffer from issues such as randomness and instabilit y of algorithm results,and the ignorance of the influence of node influence on label propagation.To address these issues,we propose an overlapping community discovery algorithm that can generate stable communities based on node influence and multi-label propagation.Based on node influence calculation,ranking and core node recognition,our algorithm realizes the effective recognition of overlapping community in complex networks,by reprocessing initial labels of neighbor nodes and updating node labels asynchronously based on balance coefficient.Experimental results on real data sets and artificial data sets show that the performance of our algorithm is better than that of other comparison algorithms and suitable for large-scale complex networks.

关 键 词：复杂网络 重叠社区 多标签传播 节点影响力 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]