基于邻居系数的聚集度可调无尺度网络建模  

Modeling of scale-free network with tunable clustering based on neighbor coefficients

作　　者：张珂[1] 黄永峰[1] 李星[1] 

机构地区：[1]清华大学电子工程系,北京100084

出　　处：《清华大学学报（自然科学版）》2008年第4期571-573,577,共4页Journal of Tsinghua University(Science and Technology)

基　　金：国家"八六三"高技术项目(2006AA01Z444)

摘　　要：针对聚集系数未涉及间接邻居连通性和无法正确描述大节点度网络节点的问题,提出聚集度的新度量-邻居系数,并基于其统计意义提出邻居系数网络模型。邻居系数从邻居演化的角度描述聚集度,定义为网络节点的间接邻居也是其直接邻居的概率,分析表明邻居系数可有效地描述各种网络节点的聚集度。邻居系数模型是通过引入局域连接这一邻居演化机制对Barabási-Albert(BA)无尺度网络模型的扩展。仿真结果表明邻居系数网络模型既具有可调的聚集度,又保持节点度的幂率分布。Clustering coefficients do not indicate the connectivity of indirect neighbors and have limited to represent high-degree node clustering. This paper presents a clustering measurement, the neighbor coefficient, with a neighbor coefficient network model based on the statistical definition of the neighbor coefficient. The neighbor coefficient describes the neighbor evolution and quantifies the likelihood that indirect neighbors are also direct neighbors. The analysis indicates that the neighbor coefficient represents the clustering of all kinds of nodes. The neighbor coefficient network model extends the Barabasi-Albert （BA） scale-free network model by introducing a local-link which is the neighbor evolution mechanism. Simulations indicate that the model has both tunable clustering and a power-law distribution of the node degrees.

关 键 词：无尺度网络 可调聚集度 聚集系数 邻居系数 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]