论文部分内容阅读
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。
A wavelet crossover technique and Volterra adaptive filter are used to predict the trend of oil price. Firstly, the original time series of oil price are analyzed by wavelet analysis, and the scale coefficients and detail coefficients of the decomposed layers are reconstructed Then the optimal delay time and embedding dimension of each time series are calculated respectively to reconstruct the phase space. Finally, the Volterra adaptive filter method is used to predict the time series of each layer to reconstruct the predicted oil price. Experiments show that this method is more accurate than direct chaotic time series global prediction and first-order local prediction, and the prediction range is larger.