Based On The GARCH Model And Forecasting The Bank Stock Index

Along with the acceleration of financial globalization,China's financial market has derived many investment opportunities,and at the same time it has also generated tremendous financial risks.The demand for investors' analyzing market capabilities has increased in the complex and volatile financial markets.They not only to applied the development of traditional knowledge of financial theory to the financial markets but also combined with more cutting-edge science and technology.Big data,cloud computing and artificial intelligence are already applied to predicting the price of stocks.However,it is necessary to use different theoretical preparations,investment strategies and analysis methods to study the research object which according to different industry environment.This article takes the closing price data of the CSI Bank Index from January 1,2015 to December 31,2017 as the research object,used the GARCH family model to establish the equations to analyze and predicts the stock price.The article first analyzes the basic statistical characteristics of the CSI Bank Index,and performs unit root test,autocorrelation test and ARCH test.The results show that there is a fat tail effect,volatility clustering effect in the CSI Bank Index Series and can be modeled by using the GARCH family model.Then,GARCH model,APARCH model,TGARCH model,standard normal,student t and GED distribution are selected for modeling separately.Finally,this article compares and selects the optimal GARCH family model and distribution for these combinations.The comparison results show that the predictive effect of APARCH-GED is best.We compare the five-day predicted value of the CSI Bank Index with the actual value,which further confirms that the model prediction effect is good.