The Study On Stock Price Manipulation Judgment Based On RBF Neural Network

After the establishment of Shanghai and Shenzhen stock exchange markets,the domestic stock market develops rapidly exposing some malfeasances,in which the stock price manipulation is most serious.The behavior can't reflect the real value of stocks,and the price discovery mechanism cannot give full play to its due functions,at the same time,it seriously affects the optimal resource allocation in the competitive market,for which a strong regulation of stock price manipulation is very necessary for the stock market in China.For an enhancement of the stock price manipulation judgment,this paper this paper analyzes the characteristics of stock price manipulation,taking stock price manipulation cases in the domestic stock market as samples.then,establishes the judging model of stock price manipulation based on RBF(Radial Basis Function)neural network.Finally,it makes related countermeasures and suggestions,which provides a new train of thought for research on the judgment of stock price manipulation in the future.As follows:The introduction is in the first chapter including the selected topic background,the literature review,the research content,the technical route,the research method and the research meaning;the second chapter expounds some core concept of stock price manipulation covering its definition,method,harm,reason,etc.;the third chapter carries out the empirical analysis taking 140 stock price manipulation cases investigated and treated by domestic securities regulatory commissions as samples,with industry characteristics analysis,the equity analysis,the comparative analysis with listed companies,the assets characteristics analysis,financial characteristics analysis,transaction characteristics analysis,etc.;the fourth chapter makes an Logistic regression analysis through SPSS taking 40 stocks as indicators of characteristics of the samples on the basis of the analysis of the characteristics,(half of them is manipulated,and the other half is not be manipulated)to screen out 4 indexes indicating the abnormal stock fluctuation,and this chapter sets up the judging model of stock price manipulation and conducts the empirical test in the Matlab environment,which shows that the detection accuracy reaches up to 90%.Therefore,the model can predicate and identify whether the stock is manipulated,which will be taken as a new perspective of future research in this field;the fifth chapter discusses related countermeasures and suggestions about ways of preventing stock price manipulation;the sixth chapter reviews the full text with a conclusion,and points out the innovation,the deficiency and the direction of the efforts in the future.