%A LIU Xiao-qian, WANG Jian, WU Guang %T The Forecasting Analysis of HS300 Index based on HAR-CVX Model of High Frequency Data %0 Journal Article %D 2017 %J Chinese Journal of Management Science %R 10.16381/j.cnki.issn1003-207x.2017.06.001 %P 1-10 %V 25 %N 6 %U {http://www.zgglkx.com/CN/abstract/article_15736.shtml} %8 2017-06-20 %X Volatility is used to measure the market risk, and its accuracy prediction has an direct significance on derivative pricing, risk management and asset allocation. How to understand and measure the market risk accurately is related to investment decision efficiency and economic operation, which is the hot problem of the government department, investors and security market. Currently, the option is in simulated trading in China, and various options will soon list orderly. To research the volatility of correspondence stock and predict it accurately is very important to option pricing and option trading. The research has great theoretical and practical significance for regulation of the financial risk and will promote the long-term health development of Chinese options market. In this paper, detailed descriptions of the characteristics of different volatilities, the theoretical of various volatility forecasting models are given. The Realized Volatility, which is based on high-frequency data, is chosen as a representative of the future real volatility. The HS300 index is chosen as the research object, and then the preliminary analysis of the basic characteristics of volatility is carried on. Combined with the stock market implied volatility, the market volatility index-CVX is chosen as the impact factor and added to HAR model. Then are used a new model called HAR-CVX model is gotten. Then several common volatility forecasting models, such as GARCH model, SV model, HAR model and the HAR-CVX model, to forecast the volatility of HS300 index.The prediction method we used is called the out-of-sample rolling time window forecast method. Finally, four loss functions and SPA test are used to evaluate the prediction results. It is found that prediction effect of realized volatility models which is based on high frequency trading data is better than the SV model and GARCH model which are based on the daily price data. In addition, Considering the CVX index, which contains the implied volatility, improves the prediction effect of the HAR model.