基于小波消噪和LS-SVM的混沌时间序列预测模型及其应用

PREDICTION MODEL OF CHAOTIC TIME SERIES BASED ON WAVELET DE-NOISING AND LS-SVM AND ITS APPLICATION

提出将小波消噪、相空间重构理论与LS-SVM相结合,实现变形监测数据的建模及预测。先对变形监测时间序列进行小波消噪;然后用C-C法求出非线性变形数据的最优嵌入维数和时间延迟参数,并对其进行相空间重构;最后采用LS-SVM对其进行建模预测,并与BP神经网络的预测结果进行了比较分析。实例表明,基于小波消噪和LS-SVM的混沌时间序列预测模型具有较好的预测效果。

On the basis of integrating the wavelet de-noising, the theory of phase space reconstruction and LS- SVM, a modeling and forecasting technique is put forward. First, the wavelet de-noising is used to pre-process data, then the best embedding dimension and time delay of nonlinear deformation data are calculated with the method called ＂C-C＂, the phase-space is reconstructed as well. At last, the modeling prediction is carried out by LS-SVM and also is compared with BP neural network. The results show that the prediction model of chaotic time series based on wavelet de-noising and LS-SVM is better.