基于选举标签传播的非重叠社区挖掘算法  

Voting-based Label Propagation Algorithm for Non-overlapping Communities Detection

作　　者：连亚飞[1] 黄发良[1,2] 汪焱[1] 潘传迪[2] LIAN Ya-fei HUANG Fa-liang WANG Yan PAN Chuan-di(Faculty of Software, Fujian Normal University, Fuzhou 350117, China Zhejiang Engineering Research Center of Intelligent Medicine, The First Affiliated Hospital of Wenzhou Medical University, Wenzhou 325000, China)

机构地区：[1]福建师范大学软件学院,福建福州350117 [2]温州医科大学附属第一医院浙江省智慧医疗工程技术研究中心,浙江温州325000

出　　处：《福建师范大学学报（自然科学版）》2017年第2期10-17,共8页Journal of Fujian Normal University：Natural Science Edition

基　　金：浙江省智慧医疗工程技术研究中心开放基金(ZH2016004);福建省教育厅资助项目(JA13077)

摘　　要：针对邻居节点选择规则过于简单的传统标签传播算法容易导致奇异解问题,从而难以适应大型复杂网络的社区挖掘,提出了基于日常生活选举模式的标签传播算法VLPNO,重新定义节点标签传播规则,使其在传播迭代过程中能依照竞选的方式自主地更新标签,进而将网络划分为由领导者和跟随者组成的社区.实验结果表明,与LPA、SLPA与BMLPA相比较,VLPNO算法能够更快速有效地发现与真实网络社区更相吻合的社区结构.Due to singular solution issue caused by simplicity of neighbor node selection rule, traditional label propagation algorithm can＇t achieve good effects in community detection of large- scale complex network. To address the issue, an algorithm named VLPNO is proposed based on voting mechanism in decision-making. Enlightened by voting mechanism in decision-making, re-defined propagation rules are presented to update labels independently In VLPNO, by which a network is divided into communities composed of leaders and followers. The experimental results show that, compared to LPA, SLPA and BMLPA, the proposed VLPNO can more fast and more effectively discover the intrinsic communities in networks.

关 键 词：标签传播算法 选举标签 社区挖掘 复杂网络 

分 类 号：TP313[自动化与计算机技术—计算机软件与理论]