Based On Hilbert-Huang Transformation High Frequency Data Analysis

A low frequency data (including linear, the steady) can be applied to some of the traditional data analysis method. In recent years, people use a new methods-Hilbert-Huang transform methodthe for high frequency data (non stationary, nonlinear). Hilbert-Huang transform is a new method of signal processing, it can according to the nature of the signal, points signal is divided into several IMF sum of component adaptively and the IMF component is irrelevant.This paper introduced the method, this paper introduces the concept of time and frequency analysis, Hilbert-Huang transform method,EMD decomposition method Hilbert spectrum analysis, Hilbert-Huang transform the problems need to be solved for EMD decomposition, arise when modal aliasing problem, proposed the improvement Hilbert-Huang transform method, improve modal aliasing phenomenon. The innovation of the duty is that in put forward a kind of method of the AR model based on EMD, and the high frequency data to EMD decomposition, get a number of IMF components, and then the IMF component were of statistical analysis, established the AR model respectively, parameter estimation then forecast,error addition will get C(t), the original data synthesis is The model prediction data and real data are compared, and the error analysis, then a conclusion will get. The empirical analysis of the application method of a stock is analyzed.Analyse the error between predicted data and real data, the result shows that this method was used for the analysis of nonlinear and the smooth high frequency data.