复杂网络中节点重要性综合评价方法研究

摘要

信息化条件下复杂网络对节点的攻击显得十分脆弱,发掘重要性节点并进行保护逐渐成为研究复杂网络理论的一个基本问题.针对以往的单个评价指标如度、介数、聚焦系数等忽略了节点在复杂网络中的综合属性,不能全面地给出评价结果的不足,提出了一种基于余弦相似度的节点重要性综合评价方法.上述方法通过计算各个节点的评价指标向量与理想化节点指标向量的余弦相似度,得出节点的综合排序结果,最终遴选出最佳节点.实验结果表明,上述方法可以较为准确地评估节点的重要性,为研究网络的抗毁性提供一定依据.%Under the condition of information,the attack on the nodes makes the complex network very fragile and the node failure may have a crucial impact on the network,therefore,exploring and protecting the importance of node has become a basic problem in the study complex network theory.According to the previous single evaluation index,such as degree,betweenness and coefficient,which ignore the comprehensive property and cannot give a comprehensive evaluation results,this paper presents a comprehensive evaluation method of complex network node based on cosine similarity.To obtain the results of comprehensive ranking of nodes and select the best node,the criteria collected based on experience were transferred as vectors,compared with the idealistic node.The experimental results show that this method can accurately evaluate the importance and provides the basis for the research on network survivability.