Based On Bayes And Likelihood Function Method To Forecast The Price Fluctuation Of Chinese Stock Market

With the advent of the information society,China's stock market is also in constant development,whether from the market,the government,or investors,have to make great progress.For the stock market,any relevant investment information will have a certain impact on the volatility of the stock market,especially the yield of the stock market.Therefore,in recent years,the research on the stock market volatility has gradually increased.For the stock market volatility,we should have rational thinking and understanding.A certain range of volatility will help to promote the development of the stock market,but if the regular volatility will cause the instability of the stock market,and will bring huge risks,which will have a certain impact on domestic investors,and further impact The next development of the stock market.For the current research on the fluctuation of China's stock market,the research object is the Shanghai and Shenzhen stock index.For investors,the volatility of the stock market is their main concern.When the stock market fluctuates,their investment strategy will also be affected.With the increase of domestic and foreign scholars' research on the stock market volatility,the relevant models are gradually improved.In order to more comprehensively understand the stock market volatility,more and more research on the prediction of stock market price is carried out.The volatility is used to reflect the size of market risk.The higher the volatility is,the greater the degree of market price fluctuation is.On the contrary,the smaller the volatility is,the smaller the degree of market price fluctuation is.Therefore,the study of volatility is of certain significance to the study of market price fluctuation.This paper selects the CSI300 index as our research object,and conducts a simple analysis of the selected GARCH model,including the development process,nature,and advantages and disadvantages of the model.The 5-minute high-frequency data of the CSI300 in 2019 are selected as the short-term sample data,and the 5-minute highfrequency data of the CSI300 from 2010 to 2019 are selected as the long-term sample data.The volatility is calculated based on the realized volatility and used as a reference for real volatility.Based on the comparison and calculation of the sample forecast of the rolling time window,we combined the realized volatility and proposed the likelihood function estimation and Bayesian estimation of the model in this paper.Finally,Akaike information criterion,Bayesian information criterion and loss function are used to test the goodness of fit of the model.The dynamic prediction performance of the two sets of data under the two estimation methods are compared.The main empirical results show that,given the data,the GARCH model based on the Bayesian Griddy-Gibbs method is better than the results of the likelihood function method in predicting the stock price and volatility of the stock market.