基于小波支持向量机的混沌时序预测算法

Chaotic Time Series Prediction Algorithm Base on Wavelet Support Vector Machine

研究混沌时间序列预测问题,混沌时间序列是一种特殊时间序列,具有高度非线性、混沌性和时变性,传统预测方法精度低。为了提高混沌时间序列预测的精度,提出一种采用小波支持向量机的混沌时间序列预测算法。首先采用相空重构对混沌时间序列进行重构,捕捉时间序列变化轨迹,然后采用小波支持向量机对非线性、混沌性变化规律进行预测,最后采用Lorenz混沌时间序列进行仿真。结果表明,相对于传统预测方法,小波支持向量机提高了混沌时间序列预测精度,降低预测误差,为混沌时间序列提供了一种有效的准确预测途径。

Research chaotic time series prediction accuracy problems. Chaotic time series is highly nonlinear and time-varying, and linear prediction method is difficult to reflect the change rule, and the accuracy of prediction is low. In order to improve the precision of predicting chaotic time series, we presented a method based on wavelet support vector machine for chaotic time series prediction algorithm. First, phase space reconstruction was used for chaotic time series reconstruction to capture the changing path of time sequence. Then, wavelet support vector machine was used to carry out the forecast. Finally, the Lorenz chaotic time sequence was used for simulation experiments. The results show that, compared with traditional forecasting methods, the wavelet support vector machine improves the prediction accuracy of chaotic time series and reduces the prediction error, which provides the nonlinear chaotic time se- ries with an effective prediction approach.