Research And Application Of Node Centrality Measurement Algorithm In Complex Networks

Complex network is the representation of complex system.Because of the large num-ber of nodes and the complex relationship between nodes,such a network is called”complex network”.The centrality of nodes in a complex network has a significant impact on the robust-ness of the network.The measurement of entropy-based node importance has become one of the hot topics in the research of complex network theory.its main purpose is to use information entropy to analyze the characteristics of complex networks and effectively predict and control the changes of complex networks and complex systems.at the same time,entropy is used to evaluate the importance of the nodes in the network to find out the most influential nodes in the network.This is an effective method to predict and control the stability of the network.There are many ways to evaluate the importance of nodes,each of which has its own advantages and limitations.For an actual network,it is one-sided to use a single index to describe the impor-tance of nodes.How to more accurately identify the importance of nodes still needs further research.This paper mainly studies some basic characteristics of complex networks and graph theo-ry,and focuses on the importance of nodes in complex networks.The main results of this paper are as follows:?1?There are many methods to evaluate the importance of nodes,each of which has its own advantages and limitations.Combined with the degree centrality DC and the betweenness centrality BC,they only considers the influence of all the neighbor nodes of node v_ion the importance of the node itself,and does not consider which part of the neighbor node set has a greater impact on the importance of node v_iitself.In this paper,the neighbor node set is divided into two parts:the related neighbor node set?M_R?and the unrelated neighbor node set(M_UR).Based on the characteristics of graph entropy,a new information entropy NBE?neighborhood betweenness entropy?and NDE?neighborhood degree entropy?are proposed.Combined with NBE and NDE,new centrality RNC?related neighbor centrality?and URNC?unrelated neighbor centrality?are proposed,and the centrality simulation is carried out experiments.The results show that the new method is feasible and effective.?2?For an network,it is one-sided to describe the importance of nodes with a single index.Combining inverse sum index?ISI??degree and betweenness of node,by setting weights for three indicators?that is,scaling parameters?,to get a new the centrality index BDI.The new index considers the centrality of the node from the three aspects of the itself,neighbors and whole networks.Compared with the single node centrality index,the new index considers more comprehensively and accurately.