Research On Node Importance Ranking Technology In Complex Networks

Ranking the importance of nodes in a complex network is helpful to analyze the structure and functional characteristics of the network,and is of great significance to maintaining the stability.In real life,the accurate identification of important nodes in the network has extremely important significance for fixed-point strikes,traffic planning,public opinion control,epidemic control,and so on.The thesis researches the node importance ranking technology from two aspects: robustness and transmission dynamics.The research content is as follows:(1)Aiming at the problem of insufficient degree centrality discrimination effect,a node importance ranking method based on local features is proposed.This method only needs to obtain the degree centrality of nodes and first-order neighbor nodes,which not only improves the distinguishing effect of node importance,but also ranks the importance of nodes more accurately.In this thesis,the effectiveness of the proposed method is verified by observing the variation graph of the proportion of independent parts and the largest connected subgraph.(2)For the problem about the influence of the overlap of the connection structure of nodes on the importance of nodes,a key node identification method based on the overlap of network topology is proposed,and the overlap of the structure between the nearest neighbor node set of nodes and the nearest neighbor node set of the first-order neighbor nodes is described by the common part.Static attack and dynamic attack experiments are performed on multiple networks.The experiment not only compare the difference between random attack and deliberate attack,but also compare the proposed method with several other deliberate attack methods.The experimental results show that the proposed method can evaluate the importance of nodes well.(3)Considering the important role of neighbor layer nodes in the process of information dissemination,a method for identifying influential nodes based on the information of the nodes themselves and neighbor layer nodes is proposed.First,the SIR model is used to simulate the influence of the nodes,and then verification experiments are carried out on multiple datasets.The verification experiments are carried out from two aspects of discrimination and accuracy.The experimental results show that the proposed method can distinguish well,and the Kendall correlation coefficient between the proposed method and the results obtained by the SIR model can reach an average of 0.9165.In all,the proposed method can show a better recognition effect than other methods.