基于拓扑稳定性的有向网络链路预测方法  

Link prediction method based on topological stability in directed network

作　　者：李治成 吉立新[1] 刘树新[1] 普黎明 Li Zhicheng;Ji Lixin;Liu Shuxin;Pu Liming(PLA Strategic Support Force Information Engineering University,Zhengzhou 450002,China)

机构地区：[1]战略支援部队信息工程大学,郑州450002

出　　处：《计算机应用研究》2020年第12期3744-3748,共5页Application Research of Computers

基　　金：国家自然科学基金资助项目。

摘　　要：作为复杂网络研究的热门方向,有向网络链路预测旨在挖掘网络中未知的连边。网络演化机制是许多链路预测方法的理论来源和依据。现有有向相似性指标计算节点间存在连边的可能性时,忽略了实际网络演化过程中节点自身拓扑稳定性和网络结构稳定性。基于此,提出了基于拓扑稳定性的预测方法。该方法首先计算趋于稳定节点对之间的相似度,然后计算预测节点对之间的拓扑稳定性。在三个衡量标准AUC、precision和排序分下,九个真实网络中的实验分析表明,提出的方法具有较高的预测精度。As an attractive area of study in complex networks,link prediction of directed networks aims to mining unknown links in networks.Network evolution mechanism is the theoretical source and basis of many link prediction methods.However,the existing similarity-based methods neglect the node’s own topological stability and network structure stability when calcula-ting the similarity between nodes.In order to solve this problem,this paper proposed a prediction method based on topological stability.Firstly,it analyzed the topological stability of nodes,and then calculated the similarity between the stable node pairs.Under the three metrics of AUC,precision and ranking score,the experimental results in nine real networks show that the proposed method has higher prediction accuracy.

关 键 词：有向网络 链路预测 拓扑稳定性 

分 类 号：TP399[自动化与计算机技术—计算机应用技术]