Signless-laplacian eigenvector centrality: A novel vital nodes identification method for complex networks
Signless-laplacian eigenvector centrality: A novel vital nodes identification method for complex networks
复制标题
DOI：
10.1016/j.patrec.2021.04.018
复制
发表时间：
2021-05-24
影响因子：
5.1
通讯作者：
Qi, Xingqin
中科院分区：
文献类型：
--
作者：
Xu, Yan;Feng, Zhidan;Qi, Xingqin
在复杂网络中识别重要节点对于理解、控制、加速或终止信息、疾病、创新、行为等传播过程具有重要意义，现有的中心性方法根据节点的相邻节点来评估节点的重要性，但往往忽略或等同处理节点关联边的影响。然而，在现实中，边缘总是扮演着不同的角色，这通常是衡量边缘中心。注意，顶点的中心性受其关联边的中心性影响，相反，边的中心性由其两个端点的中心性决定。在本文中，我们提出了一种新的方法来评估节点和边缘的中心性，同时通过构建一个相互更新的迭代框架。此外，我们将证明由该框架得到的节点中心性实际上是输入网络的signless-laplacian矩阵的主特征向量，因此我们称这种新的节点中心性方法为signless-laplacian特征向量中心性方法。我们在几个经典数据集上测试了它，并且都产生了令人满意的结果。它在未来有着广阔的应用前景。(C)2021爱思唯尔有限公司版权所有。
Identifying important and influential nodes in complex networks is crucial in understanding, controlling, accelerating or terminating spreading processes for information, diseases, innovations, behaviors, and so on. Many existing centrality methods evaluate a node's importance (or centrality) according to its neighbors, but the effects of its incident edges are always ignored or treated equally. However, in reality, edges always play different roles, which are usually measured by the edge centrality. Note that the centrality of a vertex is affected by the centralities of its incident edges, and conversely the centrality of an edge is determined by the centralities of its two endpoints. In this paper, we present a novel way to evaluate the centrality for both nodes and edges simultaneously by constructing a mutually updated iterative framework. Furthermore, we will prove that the node centralities obtained by this framework are actually the principal eigenvector of the signless-laplacian matrix of the input network, thus we call this new node centrality method as signless-laplacian eigenvector centrality method. We test it on several classical data sets and all produce satisfying results. It is expected to have a promising applications in the future. (C) 2021 Elsevier B.V. All rights reserved.