Research Of High-frequency Time Series Based On Wavelet Analysis

High frenquency time series and ultru high frenquency time series contain more information than normal series. Nowadays, the analysis of high-frequency time series is very implortant not only for economic experts, but also for people who work in financial area.What is more, in capital market, each investor has different investing way. Somebody is for long investment, and some is for short. Wavelet analysis has the multi-scale character; it's the good tool for capital market research.First, the paper discusses the persistence character of realized volatility based on wavelet analysis. Take Shanghai Stock Market for example, studies the correlation of realized volatility.Second, analyse the different Bate coeffcient of high-frequency time series. Considering the time factor, this paper not only researches the yield and risk staticly, but also solves dynamic CAPM effectively. This way can solve the problem, that chosing proper combination weight to gain more benefit or reduce the portfolio risk, on different time scaled.The last, discuss the forecasting of high-frequency data based on wavelet analysis. In this paper, we dispose the stock market's high-frequency data into two parts: high-frequency series and low-frequency series by the method of multi-resolution wavelet decomposition and reconstruction. The result shows that this method for forecasting is good and it provides a new method on forecasting for high-frequency data.At last, the paper reviews the research of this paper and points out its'research trend.The research is sponsored by National Natural Science Foundation of China: Research on Long Run Eqilibrium in Multivariate Moments Series and Avoiding Tactics of Dynamic Financial Risk (No. 70471050).