Research Of Key Edges Recognition Algorithm In Complex Networks

Complex networks are a powerful tool to study the interaction of real system,edges directly represent the connection between nodes and they play an important part of the network.The role of edge in network is heterogeneous,while key edges play a decisive role in ensuring the smooth flow of information and the integrity of network structure.Algorithm of mining fast and accurate key edges recognition has attracted extensive attention of many scholars,at present,they have proposed a variety of key edges recognition algorithms from different angles.Identifying key edges is not only of theoretical significance,and it has practical application value in guiding network protection and control strategy.In this paper,we mainly focus on the problem that the key edges recognition algorithm based on network structure ignores the interaction between edges and the high time complexity of key edges recognition in large-scale network,main research as follows:(1)At present,key edges recognition algorithms mainly rely on the topological structure of the nodes,but ignore the deep structure information between edges,which makes the accuracy of key edges recognition is not quite satisfied.Therefore,this paper studies the Weighted K-Shell(WKS)algorithm for identifying key edges on edge structured networks.WKS algorithm weights the importance of remaining nodes and removed nodes,which affect the importance of nodes.It solves the problem that only considering a single factor leads to low accuracy of key nodes identification.Finally,in order to verify the accuracy of the algorithm to identify the key nodes in the edge structure network,the SIR epidemic model experiments are carried out with WKS algorithm and other four key node identification algorithms in four real networks,and the propagation speed and propagation range of the key nodes identified by WKS algorithm are analyzed.The experimental results show that WKS algorithm can effectively and accurately identify key edges of the original network.(2)The structure of large-scale network is complex and the number of edges is huge,but the number of decisive edges only accounts for a small part of it,so not all edges need to be analyzed when identifying key edges.In order to reduce the unnecessary time consumption in key edges recognition,the redundant edges which have little effect on the network structure and function are compressed.On the compressed network,key edges recognition algorithm based on redundant edge compression is studied.The algorithm uses Connectivity Efficiency(CE)to quantify and sort the importance of edges,and finally identifies the key edges.In order to verify the efficiency and rapidity of the proposed algorithm,experiments are carried out on six largescale networks with four other key edge recognition algorithms.The experimental results show that the proposed algorithm can recognize key edges accurately and quickly.Starting from the network structure,this paper studies two key edges recognition algorithms,which provide a reliable theoretical basis for in-depth analysis of the interaction between edges and improving the real-time performance of key edge recognition in large-scale networks.