Chinese Stock Market Volatility Forecasting Model And Its Application

In the modern financial market theory, financial assets volatility describes is one of the most important elements. The description of the volatility of financial assets mainly in the following two types:①the Implied Volatility Model, this description method based on the data of options pricing.②the Historical Volatility Models, the construction of this kind of volatility models derived from historical income data. In the History Volatility Models, the most representative are those, that ARCH Model, GARCH Model, SV Model and RV Model. Traditionally, the measure with day income rate of square as day fluctuations rate will be faced on very serious of measurement errors and noise, and the use of high frequency trading days based on the data of the "has realized" volatility as the volatility measure, will greatly reduced these measurement errors and noise on real volatility process of effect, and as the high frequency of income increased, the measurement error will become smaller and smaller.We can see from above, the characterization of financial market volatility can have a variety of different model description methods, however, Which method is the most suitable for China’s stock market fluctuation of the actual condition and risk characteristics?In recent years, domestic scholars on the volatility of stock price index in China have also carried out a number of useful explorations of the model. To date, the domestic existing research still exist in some need for improvement and further research problems. In order to explore Chinese stock market fluctuation characteristic comprehensively, and choose the optimal volatility forecasting model, empirical analysis of this article is to the CSI300as an object, and this paper make efforts to study the following issues:Most research comparison has based on low frequency data of GARCH family model and SV model forecast results only, this paper attempts to conduct a comprehensive comparison of prediction results, witch is based on the high frequency data "realized" volatility model (RV Model) and based on low-frequency data historical volatility model (HV Model). An empirical test of different volatility model prediction accuracy, witch is to increase the stability and utility of the conclusions.