一种基于局部特征的节点重要性排序方法  

A Ranking Method of Node Importance Based on Local Features

作　　者：朱敬成 王伦文 吴涛[1] ZHU Jing-cheng;WANG Lun-wen;WU Tao(College of Electronic Engineering,National University of Defense Technology,Hefei Anhui 230037,China)

机构地区：[1]国防科技大学电子对抗学院,安徽合肥230037

出　　处：《计算机仿真》2022年第11期416-421,共6页Computer Simulation

摘　　要：复杂网络中节点重要性排序对于研究网络的鲁棒性具有十分重大的意义,有助于提高网络的抗毁性,使得网络处于稳定的状态。而在大规模网络的节点重要性识别过程中,获取全局信息计算复杂度较高。度中心性在众多方法中最简单且复杂度最低,但存在无法区分相同值节点的重要性的问题,考虑到节点重要性不仅和节点本身有关,还与节点的邻居信息有一定的关联度,所以引入了一阶邻居对节点重要性的影响,提出了一种简单且有效的局部特征排序方法,提高了节点重要性排序的分辨率和精度。在六个真实网络和三个人工网络上进行验证实验,以独立部分比例和最大连通子图比例为评价标准,实验结果表明,所提方法与度中心性DC、半局部中心性SLC、基于节点及其邻居度的WL方法、K-shell分解方法、DKN方法以及基于节点和边的NL方法相比更加有效。The importance ranking of nodes in complex networks is of great significance to the study of network robustness,which helps to improve the invulnerability of the network and make the network in a stable state.However,in the process of node importance identification in large-scale networks,the computational complexity of obtaining global information is relatively high.Degree centrality is the simplest and the lowest complexity in many methods,but it is difficult to distinguish the importance of nodes with the same value.Considering that the importance of nodes is not only related to the node itself,but also has a certain degree of correlation with the neighbor information of nodes,this paper introduced the influence of first-order neighbors on the importance of nodes,and proposed a simple and effective local feature ranking method to improve the performance of the algorithm.The resolution and precision of node importance ranking were also discussed.Experimental results on six real networks and three artificial networks show that the proposed method is more effective than degree centered DC,semi local centrality SLC,WL method based on node and its neighbor degree,K-shell decomposition method,DKN method and NL method based on node and edge.

关 键 词：复杂网络 鲁棒性 局部特征 节点重要性 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]