Analysis Of The Impacted Factors And Prediction On The Price Of Stock

Since the establishment of Shanghai Securities Exchange, the prediction of stock price has become the focus of attention. Common prediction models are becoming more and more, such as Markov model, Grey System model, ARMA/ARIMA model, ARFIMA model, Neural Network model, Chaotic Dynamics model, Non-parametric kernel regression estimation method, GARCH model and so on. But the stock market is a very complex nonlinear system, often affected by the international environment, national policy, economic siruation, political situation, a variety of factors, and investors'psychology. In the past research results, most are based on historical stock prices in order to predict future prices, with little consideration on its factors.In this paper, the methods of multivariate statistical are used to analysis the correlation of the stock price and its factors, main factors will be introduced to FAR model, FAR forecast model with exogenous variables of the stock price is established. The content of the thesis are as follows:(1) The canonical correlation analysis model of stock price with its impact factors is established, then canonical correlation variables are got, and the correlation coefficients of canonical correlation variables are significant test. The result shows that there is a strong correlation between them. Last the main influence factors of stock price are found.(2) The grey canonical correlation analysis model of stock price with its impact factors is established, then grey canonical correlation variables are got, and the correlation coefficients of grey canonical correlation variables are significant test. Then the main influence factors of stock price are found. At last, compared with the canonical correlation analysis model. The result shows that the grey canonical correlation analysis model is more applicable to the sample data with time significance.(3) Based on the above analysis, according to the theory of FAR model and local linear estimation, with MATLAB programming, the function coefficient autoregressive model with exogenous variables of the monthly average data of Shanghai and Shenzhen 300 index and monthly data of its impacted factor (money supply) is established. Not only the influence of lagged endogenous variables is considered, but also the influence of exogenous variables is considered in this model. Lastly, this model is used to predict the stock price of a few months later and the forecast results are accurately. The results show that this model has a good prospect of application to predict the stock price.