Optimization Of CSI300Index HAR Model’s Structure Using Genetic Algorithm

Volatility research has been an important issue of the Financial Research. The volatility is the main features of the financial asset price changes, choose a suitable estimator to model volatility and forecast its future variation, it provide strong support for the pricing of financial assets, portfolio identification and risk management. As the lower cost of information storage, the financial research gradually shifted from the low-frequency field to high-frequency field, high-frequency data contains more information, it will make the volatility forecast more complex.Realized volatility is based on the development of the high-frequency data, heterogeneous autoregressive model (HAR model) is a model of realized volatility, as it simple to calculate and accurately to estimate, so that gradually developed in the field of study of the volatility.The text details the development of the the realized volatility HAR model, and research HAR model applicability of the Chinese market. Use the HAR model regression analysis CSI300Index,which is able to represent the characteristics of the Chinese market. Discussing under the (1,5,22) time scales, the forecast capability to the CSI300Index. Then using genetic algorithms to optimize the time scale of the model, geting a new time scale can reflec the Chinese market investor behavior characteristics better. Provides references and suggestions for the development of HAR models in Chinese market as well as the CSI300Index volatility forecast.