Research On Important Node Identification Methods In Temporal Complex Network

As a hot topic of research in complex network,important node identification plays an important role in placing advertisements,breaking the spread of infectious diseases and controlling the spread of public opinion.Temporal network is an extension of the study of complex network.Compared with static network,temporal network can portray the real system more accurately by considering the temporal property,and has received extensive attention from many scholars in recent years.In this thesis,we focus on the important node identification method in the temporal network,and the major contents of the study are as follows.(1)An identification method of important nodes in temporal network based on degree and K-shell is proposed.The degree and K-shell algorithms are two commonly used methods for identifying important nodes,but they evaluate the importance of nodes only from a single aspect,which leads to a lack of comprehensiveness when evaluating the nodes’ importance.To deal with this problem,this thesis studies the nodes’ importance in the network from a global perspective by combining two metrics,degree and Ks value of nodes.Firstly,we slice the temporal network;secondly,we calculate the degree and Ks value of nodes on each slice,and determine the weight of degree and Ks value by entropy weight method;then we compute the weight factor of degree and Ks value according to their weights,and then compute the nodes’ importance on the current slice;finally,we calculate the nodes’ importance in the temporal network by incorporating the average temporal aggregation model,and pick the accuracy and distinction as the evaluation criteria,and perform experiment on various datasets to validate the effect of the method in this thesis.(2)A temporal K-shell Iteration number algorithm based on gray correlation analysis is proposed.The K-shell algorithm has the merit of low time complexity,but the algorithm ignores the number of iterations in the process of removing nodes,resulting in a large number of nodes in the network being given the same importance,and because the algorithm does not consider the temporal property,it cannot accurately portray the real system.To address this problem,this thesis proposes a weighted Kshell iteration number algorithm,and combine with gray correlation analysis proposing temporal K-shell iteration number algorithm based on gray correlation analysis.Firstly,we slice the temporal network;secondly,we use the weighted K-shell iteration number algorithm to calculate the importance of nodes on each slice and construct the sample matrix;then we compute the importance of nodes in the temporal network by incorporating the gray correlation analysis;finally,we conduct relevant experiments on different datasets and verify the effectiveness of the method based on two evaluation criteria: accuracy and distinction.