Simulating model based on the data of the closing price of Shanghai Composite Index and New Composite Index, aiming to explore and forecast the trend of Shangfhai Composite Index and New Composite Index.Firstly, analyze the market of Shanghai Composite Index and New Coinposite Index on the Shanghai Stock Exchange and found that the closing price of Shanghai Composite Index and New Composite Index present on non-stationary characteristics,deciding to adopt a logarithmic rate of return to analyze.After twice Exception Handling of the rate of return series of New Composite Index, through a series of statistical tests,estimation and finally establish ARMA-GARCH model under the different assumptions(residual distribution:normality, t-distribution, GED-distribution) comparison of fitting results under different distributions and concluded that:First, the rate of return series of Shanghai Composite Index and Operated New Coinposite Index both can be established the ARMA(2,2)-GARCH(1,1) model;Second,when the error distribution for the GED-distribution, the effect of predicting on the rate of return series of Shanghai Composite Index and Operated New Composite Index is good. Third, through twice Exception Handling of the rate of return series of New Composite Index, after each treatment, making the results of the short-term prediction model showed significant advantages. |