小波分频技术和混沌时间序列在国际石油价格预测中的应用

提供了一种小波分频技术结合Volterra自适应滤波器的预测石油价格趋势的方法,先对原始的石 油价格时间序列进行小波分频分析,将分解后的各层尺度系数和细节系数重构各层的时间序列, 然后分别计算各层时间序列的最佳延迟时间和嵌入维数来重构相空间,最终用Volterra自适应滤波器法预测各层时间序列, 重构成预测油价.实验证明该方法比直接混沌时间序列全局预测和一阶局域预测的精度更高,可预测范围更大.

World oil price forecasting based on wavelet analyze and chaotic time series technology

A new algorithm for worm oil price chaotic time series prediction based on wavelet analyze and Volterra self adaptive filter method is presented. Firstly, the original oil price time series is decomposed as the measurement coefficients and wavelet coefficients by utilizing the stationary wavelet transform.Secondly, the coefficients are predicted with a Volterra adaptive filter in their reconstituted phase spaces based on the chaotic time series method. Finally the predictions of the coefficients are acquired by the inverse wavelet transform. The result shows that the proposed method can capture the dynamics of the nonlinear systems series effectively.