The Estimate Research Of Chinese Stock Returns Based On Artificial Nerve Network

As for stock returns estimate research, is a very huge topic, at dissimilarity purpose leading bottom will have variety dissimilarity of research angle. The main purpose of this paper is according to the investor to carry on an estimate analysis to the stock returns, writer hope pass to scoop out the history data, use the method of artificial nerve network to imitate analysis the history data, to carry on some market phenomenon and variety explanation.Firstly, this paper introduces the research presently condition of stock returns in domestic and international, and simple analysis the development stock returns and estimation; Then introduce stock rate of return and estimation of research method, which lay theories foundation for empirical analysis.Secondly, analysis the impact factor of the stock returns from the theories. In the first, analyze the F-F "three factor models" analysis and estimate to stock returns, the "three factor models" is mainly description market factor, scale factor and the rights influence of the factor upon the stock returns; Later analyze the influence to stock returns for trade amount, policy, domestic and international economy.Thirdly, make use of research method of the artificial nerve network, take "three factor models" of Fama and French as foundation, and on this foundation improve the model, join trade amount and total risk coefficient as change quantity, namelyβrisk coefficient,σ2 total risk coefficient, scale, income price ratio and trade amount as change quantity. In this paper analysis five change quantity, all date come from the CSMAR database, and use EXCEL of the Microsoft carry on those data sorting and first step calculation, then ducting in the nerve network within MATLAB software tool box, and establish the network carry on empirical analysis, then carry on an estimate research each impact factor relate to stock returns, thus for stock returns estimate provided research way of thinking.Lastly, by collecting the heavy stocks of each profession, make use of MATLAB, get conclusion pass empirical analysis conclusion: Income price ratio and trade amount is biggest to stock returns influence, total risk coefficient and scale time it, theβcoefficient is minimum. And give the related investment suggestion of the investor and take charge of a section by conclusion. Suggestion investor according to each impact factor of size selection appropriate investment object, suggestion take charge of a section strengthen a management, take period as investor exaltation investment returns, for investor actual investment provide theories instruction.