| With the development of information technology,various network systems have emerged and developed rapidly.Finding key nodes in these network systems is of great significance to the normal operation of the network system.For example,in the process of disease transmission,quickly identifying super communicants and reducing their contact with susceptible populations can play a significant role in the prevention and control of disease transmission.Therefore,this thesis applies the combination of local information and information entropy of nodes to the model for identifying influential nodes,and proposes a corresponding improved model.The main research work is as follows:1.Aiming at the current problem of single evaluation index for node spreading capability,this thesis proposes a new index for evaluating node spreading capability-spreading entropy.The node relationships in a network are complex,and it is difficult to accurately determine key nodes based solely on domain or global information.Therefore,based on the probability characteristics of information spread in the network,this thesis proposes a new indicator to detect the spreading ability of nodes in the network based on node information entropy.Experimental results show that this index can quickly identify nodes with strong spreading ability.2.In order to solve the problem that the K-shell centrality method has low differentiation of node importance for the same K-shell,this thesis proposes a Kshell centrality method based on entropy improvement.In this thesis,we first calculate the ratio of the target node’s K-shell value to the total K-shell value of neighbor nodes to determine the K-shell contribution of the target node,and then proceed from the overall network situation,combining the information entropy of the node to further amplify(or reduce)the influence of the node.Experiments show that this method can quickly identify influential nodes in small networks.3.Aiming at the poor stability and adaptability of existing gravity models,this thesis proposes an entropy based gravity model.The model considers both local information(local importance)and global information(propagation probability)of nodes,and uses spreading entropy indicators to enhance the functionality and applicability of the model.Experimental results show that it can exert performance advantages in larger networks. |
