| We live in a world dominated by complex systems.With the rapid development of information technology,on the one hand,the basic structure of complex systems in human society-complex networks,has become more and more complex,on the other hand,what people can obtain,The network data sets that need to be processed and can be processed have become rich,diverse and large in scale.At the same time,the world is more and more closely connected,and human global systemic risks appear more and more frequently.The bridge in the network structure plays an important role in detecting the community structure,studying the law of information dissemination,the epidemic research of infectious diseases,and the reliability of the network.However,among a large number of researches on complex networks,only a few are specifically aimed at bridges,and bridge identification is still a problem that has not yet been effectively solved.Based on the research results of predecessors,this paper constructs several characteristic indexes of the local topological structure of the network based on edges for the purpose of identifying bridges in the network.According to the experimental results on real network data,this article has the following findings: 1.Using the features constructed in this article,based on supervised machine learning,bridges in the network can be accurately identified;2.Under the experimental conditions of this article,bridge identification is only One feature is needed to achieve a recognition effect equivalent to multiple features;3.Different network data has different optimal features;4.The model’s recognition effect on bridges is not sensitive to the proportion of the training set to the total data size,that is In the same network,the local topological structure is highly consistent;4.The first three most important features on all data are almost all generated from the four features of local＿load,local＿similarity,min＿core＿degree,and cosine＿similarity1,reflecting the community structure in the real network There is a certain commonality. |
