Importance Calculation of Nodes in Complex Network Based on Improved TOPSIS

摘要

It is of great significance to discover and protect important nodes in complex networks. In order to evaluate the importance of nodes in complex networks scientifically and overcome the one-sided shortcomings of single judgment methods such as complex network topology, node removal and node constringency method, this paper propose a scientific method for evaluating node importance. Firstly, we seperately evaluate the network topology and evaluate the importance of nodes by destroying the network structure to calculate the following 4 kinds of index: the degree value, closeness centrality, betweenness centrality and the degree of network efficiency reduction after removing nodes. Secondly, using the improve TOPSIS (technique for order preference by similarity to an ideal solution) algorithm to fuse all index, calculate the relative approach degree and rank the importance of each node. Finally, evaluating the method using the "ARPA network" and "Italian network", the result of the experiment confirmed that this method is more accurate than the single node importance evaluation method and verified its feasibility and scientificity.