Analyzing The Potential Influnence Of A Stock Market Based On Link Prediction Method

In recent years,a large number of researchers mainly through the complex network topology and centripetal approach to describe the basic characteristics of complex networks.In the network information society,the complex network has been widely used in all walks of life,in which the financial securities market is an important network of complex discussion direction,making the complex network link prediction was promoted to the main research.In the financial securities market,the volatility of the stock price must have a certain risk.This will not only break the market to bring the equilibrium value,and may even lead to the complete loss of market performance.Therefore,this paper will use the correlation coefficient method and the threshold method to construct the stock network market,using the local structure center to analyze the potential impact of the stock market,using link forecast similarity indicators to measure the potential impact of the stock in the future the degree of possibility of interconnection.The main work of this thesis is:Firstly,we are mainly from the center of the topology to measure the degree of important impact of the node.We give the advantages and disadvantages of them through the central value of the local metric and the global metric.Therefore,this paper uses the local structure center method to analyze the influence of the stock in the A-share market.In this paper,we will use the local structure center method to analyze the influence of stock in A-share market.It's accuracy is far higher than the local measure,and the computational complexity is far below the global metric.Secondly,we collect the closing price data of the daily A shares to calculate the logarithm correlation coefficient of the stock price,which briefly reflects the correlation between the stocks.Then we use matrix measurement and threshold method to analyze the stock market model.Therefore we study that the stock market model of the threshold 0.89 is more appropriate.We using the local structure center method to analyze the potential influence stock in the stock market when the different thresholds.Finally,we know that the greater the similarity between the stock nodes,the greater the likelihood of the connection between them.We use the control variable method to compare the local information similarity with the AUC as the evaluation criterion.Then we obtained that the resource allocation(RA)similarity index is the best ideal similarity index.Based on the dynamic evolution of the network in the similarity index betweenthe RA index and the local path,we can see that the optimal similarity index is constantly changing.The study shows that the specific network form makes use of the RA index and the Katz index to analyze the stock in the A shares,and the possibility of the connection between the A shares and the stocks with the potential influence of the nonferrous metals,so that the investors can reach the circumvention Risk effect.