Study On Chaotic Time Series Based On Wavelet And Support Vector Machine

Chaos is an irregular behaviour which is a wide existent phenomenon. Chaos is a complicated behaviour which is from a deterministic nonlinear dynamical system. Since chaotic systems are frequently encountered in various fields, research on chaotic system thus has great practical significance. The data we gained sometimes are chaotic time series of some variable in the practical chaotic systems. Chaos is a deterministic system determined by the nonlinear dynamical mechanism. There is a deterministic rule in the interior of the chaotic system which is seemed as a random move. Chaotic time series can be predicted in the short term. But, there are some problems in research on chaotic time series. And the study hasn't reached perfection.So we launch the deep research in view of the chaotic time series analysis and forecast technique's improvementIn this paper we first discuss the theory of chaotic time series and phase space reconstruction, the estimation of phase space reconstruction's parameter and methods of choosing embedding dimension and delay time.Chaotic time series are extremely sensitive to its initial conditions. It is very difficult to analyse chaotic time series. We mostly utilized wavelet transform and support vector machine in this paper when we analysed the chaotic time series. Wavelet transform is good at featute extraction and eliminating noise. And support Vector Machine is a kind of novel machine.Because it has quite perfect theoretical properties and good learning performance, and can solve some practical problems with low sample size, non-linearity, high-dimension of feature space, local minimization and classification ability. We integrate support vector machine and wavelet transform and improve the precision of forecast when study the chaotic time series.Finally, for real chaotic time series used in our research, this method reveals the efficiency and accuracy.