Study On Influence Ranking And Influence Maximization Of Nodes In Complex Network

With the rapid development of Internet and the explosive growth of network data,the complex network has become the research focus of experts and scholars at home and abroad.In recent years,during the research process of complex networks,a large amount of real network data has been collected and the characteristics of complex networks in different fields have been summarized.The researchers have realized that the research about the influence maximization of nodes in complex networks has a very important significance for the commercial promotion,information monitoring and control of disease transmission in the social development.Based on a large number of literatures,the basic theories related to the complex network field were studied in this paper,and the ranking problem of the k-shell decomposition algorithm and structural holes features was deeply analyzed.The deficiency of the influence maximization of a selected best seed node was also analyzed in this paper.Combined with the current research status and existing problems,the following two algorithms were proposed in this paper.First of all,aiming at the coarse-grained division problem of the k-shell decomposition algorithm,a kind of node influence rank algorithm based on k-shell and structural holes features was proposed in this paper.The algorithm considers the existence of pseudo core nodes in a network where the structural holes features form and the reality distribution in the network,and combines the function of the “structural holes” restraint coefficient on the local attributes of the nodes and the effect of the attenuation function on the pseudo core nodes,a correlation coefficient index of the node influence ranking is obtained and therefore the accuracy of the node influence ranking is identified.Secondly,based on the in-depth study and analysis of the characteristics of influence maximization of the heuristic VoteRank algorithm,a kind of heuristic influence maximization algorithm based on two-level voting was proposed.Considering the influence overlapping between node-sets,the algorithm takes fulladvantage of the voting contribution degree of the sub neighbor nodes,selects the best seed node set and adopts an independent cascade model.And the effectiveness of the proposed algorithm is evaluated according to the measure indicators that infection ratio and infection probability.Finally,three real network datasets were selected to perform simulation experiments on the SIR model and the independent cascaded model,and compared with several classical algorithms.In this way,the experimental results were analyzed and conclusions were drawn.