The Prediction Research Of Stock Price Based On Varying Coefficient Autoregressive Model

With the development of the social and economic market and the deepening of the reform,the development of the financial market is more mature and diversified.As one of the important components of the financial market,the stock market is an intricate power system.Therefore analyzing the relativity of stock market and predicting the changing trend of stock price become an important subject in financial and academic circles.In this paper,the Copula function is used to study the transformation relationship between stock prices.Based on thevarying coefficient autoregressive model,the variation regular of Hushen 300 index is predicted.The main work of the paper is as follows:First of all,in order to explore the dynamic relationships of stock market and between the stock market and the money market,The self-changing law of Hushen 300 index is successfully captured by establishing a univariate varying coefficient autoregressive model(FCAR5(3)).Compared with the FCAR5(3)model,the prediction accuracy of the varying coefficient autoregressive model(FCARSE5(3,2)),which had Shibor as the exogenous variable,is improved.Shibor has a corresponding impact on the closing price of the stock:which is more obvious when the Hushen 300 index falls.Secondly,in order to analyze the relationship between the Hushen 300 index and Zhongzhen 500 index,Gaussian-Copula and t-Copula functions are used to simulate the joint distribution of the two indexes respectively.The empirical results show a strong nonlinear correlation between the Hushen 300 index and the Zhongzheng 500 index,and the two stocks have the same downward trend.The t-Copula function is more effective in simulating the joint distribution of the two indexes.Finally,in order to explore the possibility of combining Copula function with varying coefficient autoregressive model,the probability density of two random variables calculated by Copula function is regarded as smooth variable of varying coefficient autoregressive model,thought it to achieve the effect of improving the model.In the empirical part,we established the pre-and improved varying coefficient autoregressive models with the Zhongzheng 500 index as the exogenous variable.Compared with the pre-improved model,the prediction effect of the improved varying coefficient autoregressive model(FCARSE3(2,5))with t-Copula function is obviously improved.It is indicated that the post-improved varying coefficient autoregressive model has application value in studying and predicting the changing regulation of stock market.