基于节点互动力的局部社团发现算法  

Local community detection algorithm using interactive force of node

作　　者：王天宏[1] 武星[2] 兰旺森[1] 张慧芳[1] 

机构地区：[1]忻州师范学院数学系,山西忻州034000 [2]上海大学计算机工程与科学学院,上海200072

出　　处：《计算机工程与设计》2016年第5期1291-1296,共6页Computer Engineering and Design

基　　金：高等学校博士学科点专项科研基金项目(20123108120027);忻州师范学院重点建设学科基金项目(2012);忻州师范学院青年基金项目(QN201317)

摘　　要：传统的局部社团检测算法虽然在检测社团质量上很出色,但往往依赖于起始节点的选择,在吸收新成员规则上过于严格或者预设参数难于获得,为此提出一种基于节点互动力的局部社团发现算法。节点互动力是网络成员间引力关系衡量标准,能真正反映节点间或节点与社团间互作用的强弱。以网络中局部度最大节点作为暂时社团种子,计算所有节点和社团的互动力,以互动力为标准,选取最大互动力节点作为待加入成员,直至全部成员完成社团划分。基于已知真实网络和人工网络的实验结果验证了该算法的有效性。Traditional local communities detection algorithms generally rely on the choice of the start node and put strict policy on agglomerating new vertices.Moreover,they have predefined parameters which are difficult to obtain.A local communities detection algorithm based on the node interactive force was proposed.The node interactive force which is a metric for the relationship between members of the network,it can truly reflect the interaction between nodes or that between nodes and the community＇s strength.Network nodes with the local largest degree were regarded as temporary community seeds,the interactive forces of all the nodes and communities were computed as standards,the maximum interaction force nodes were selected as a community member until all the members of the network were divided.The test based on the known real network and the synthetic network verified the effectiveness of the algorithm.

关 键 词：复杂网络 社团发现 互动力 模块度 人工合成网络 

分 类 号：TP393[自动化与计算机技术—计算机应用技术] TP311[自动化与计算机技术—计算机科学与技术]