Heteroscedastic variance is common in the financial time series.It's the core parameter in CAPM,APT,option pricing model,and plays a key role in the VaR system,so it's important to estimate and forecast the volatility.This paper propose the significance of the research at first,then review the existing research and compare the usually used GARCH model and Stochastic Volatility model.It turns out that the SV model performs better in the simulation of volatility than the GARCH model.Then we continue to compare the SV-N model and the SV-T which is so called fat-tailed model,and it turns out that the SV-T model performs better than the SV-N model in the simulation of the Chinese stock market volatility.AT last,we use the SV-T model to test if the index future and the security reform have influences in the volatility of the Chinese stock market,which has the realistic meaning in the Chinese stock market.This paper gives the direction of the research in the future. |