The Estimation Of High-frequency Data Volatility Based On Hilbert-Huang Transform

High-frequency financial data volatility estimates in asset pricing and financial engineering plays an important role. A good trading strategy in risk assessment, reliable method for the recent volatility estimation and short-term estimation is very important. Generally, with the increase of frequency data, estimate the difficulty of the volatility also will increase, for example, in the risk assessment or the financial market, stock index and the foreign exchange rate with the increase of frequency, the nature of the random process will change, self-similarity move along with the price for a long time interval and destroyed, so accurate estimates of the short-term volatility is more and more important.The characteristics of this paper mainly from two aspects of fractal characteristics and volatility of China’s stock market. First of all, according to the Shanghai and shenzhen 300 index closing price of the logarithmic return rate, the empirical mode decomposition, obtained a series of intrisic mode function(IMF: Intrinsic Mode Function); Secondly, the empirical research on the intrinsic mode function using R/S(R/S: Rescaled Range Analysis) method in fractal theory, obtain its Hurst moving average, then revealed the China’s stock market has obvious long memory in 2010-2011 years; Finally, on the high frequency intrinsic mode functions of Hilbert transform and construct a volatility estimation, also shows the volatility estimation results, and the empirical result shows that Hilbert-Huang transform is one of the reliable precision of volatility estimates method.