Research On Vital Nodes Identification In Complex Network Based On Improved K-shell Algorithm

Nowadays,human beings have gradually entered the era of science and technology,and the connection between things become closer and more complicated.In real life,network is everywhere,common ones include social networks,transportation networks,power networks,etc.All these networks can be regarded as complex networks.Complex network theory research has gradually matured,studying the internal laws of complex networks has become the main research direction,vital nodes identification is a hot issue of it.The importance of evaluating nodes is of great significance for studying the structure of networks and the dissemination of information.The K-shell algorithm mainly determines the importance of the node based on the position of the node,which causes a large number of nodes to have the same importance.In view of the coarse granularity of the network decomposition,this paper proposes a hybrid K-shell vital nodes identification algorithm based on entropy method.Adding vertex strength to improve the network decomposition standard,and using the entropy method to objectively measure the degree and vertex strength,it can effectively avoid the occurrence of a large number of nodes obtaining same score.In addition,in order to comprehensively evaluate the importance of nodes,the information of local attributes and neighbor nodes is added when calculating the final score of the node,so that the vital nodes identified by the algorithm are more reasonable.The K-shell decomposition algorithm obtains the core value of the node by decomposing the network continuously.During the deletion process,the information of the nodes in the original network will be lost.In order to link the information of the nodes in the previous network and the network after the node deletion,this paper proposes a multiple factor vital nodes identification method based on network efficiency.After the node is deleted,the efficiency of the network would change,so use relative reduction of network efficiency can further improve the ability of identifying important nodes.Finally,the experimental results also prove that the method proposed in this paper is more reasonable for evaluating the importance of nodes,and can accurately identify vital nodes in the network.