基于资源传输匹配度的复杂网络链路预测方法  

Link prediction method based on matching degree of resource transmission for complex network

作　　者：刘树新 李星[1,2] 陈鸿昶 王凯[1,2] LIU Shuxin;LI Xing;CHEN Hongchang;WANG Kai(Information Technology Institute,Information Engineering University,Zhengzhou 450002,China;National Digital Switching System&Engineering Technology Research Center,Zhengzhou 450002,China)

机构地区：[1]信息工程大学信息技术研究所,河南郑州450002 [2]国家数字交换系统工程技术研究中心,河南郑州450002

出　　处：《通信学报》2020年第6期70-79,共10页Journal on Communications

基　　金：国家自然科学基金资助项目(No.61803384)。

摘　　要：为了解决基于资源传输的链路预测方法忽略节点间匹配度对资源传输过程影响的问题,提出了一种基于资源传输匹配度的复杂网络链路预测方法。首先,对资源传输路径上的2个端点进行详细分析,提出任意节点间匹配度的量化方法;然后,为了刻画匹配度对于节点间资源传输过程的影响程度,定义了资源传输匹配度;最后,基于资源传输匹配度,考虑节点间双向传输的资源量,提出资源传输匹配度指标。在9个实际网络数据集上的实验测试表明,相比其他基于相似性指标,所提方法在AUC和Precision衡量标准下能够取得更好的效果。In order to solve the problem that many existing resource-transmission-based methods ignore the important influence of the matching degree of two endpoints on resource transmission,a link prediction method was proposed based on matching degree of resource transmission for complex networks.Firstly,by analyzing the two endpoints on the resource transmission path in detail,the method of quantifying the matching degree between two nodes was proposed.Then,in order to describe the influence of matching degree on resource transmission process between nodes,the matching degree of resource transmission was defined.Finally,based on the matching degree of resource transmission,a resource transmission matching index was proposed considering the resource amount of bidirectional transmission between nodes.The experimental results of nine datasets show that compared with other similarity indices,the proposed index can achieve higher prediction accuracy under the AUC and Precision metrics.

关 键 词：复杂网络 链路预测 资源传输 匹配度 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]