Volatility Model Study Based On Data Of Different Frequencies

The stock market is both complicated and varied, the volatility of which has become one of the core objection studied in this field. As the emerging financial market, the Shanghai and Shenzhen stock market have been stable since the imple-mentation of price limit, but the market still need to be improved further. Volatility is an important indicator to measure quality and efficiency of the stock market. A certain range of fluctuation is advantageous to active market and make it sustainable development, while too frequent and violent fluctuation will backfire. Therefore, It is great significant to investigate the stock market volatility from the view of the standardization stock market as well as the investor’s rational judgment.Based on the most representative of the Shanghai composite index and Shen-zhen composite index as the research object,the latest historical data is used to analyze our country’s stock market volatility in recent years, then the daily data,5minutes data and1minute data of each stock are selected to establish two kinds of average model and six kinds of GARCH variance model. Researching the corre-sponding volatility and drawing the following conclusions:On the one hand, in recent years, the most suitable research models of average earnings volatility of Shanghai and Shenzhen stock market in our country are the ARFIMA model, variance model for the EG ARCH model.On the other hand, a different optimal volatility model is obtained when the frequencies of earnings sequence is different. The higher the frequency, the better the fitting effect and the prediction effect of the volatility model. In this paper, the results of the study of the volatility of1minute data model is the most ideal.For one thing, with the improvement of the Chinese stock market, stock market has gradually beocme stable, the returns of which have long memory time series. For another, both the ARFIMA-EG ARCH model consider the autoregressive heteroscedasticity, the leverage effect, the long memory of earnings sequence and the risk factor. To the end, high frequency of the data means more comprehensive information effect about the stock market and also more help fit and forecast of the volatility model.