为提高对非平稳时间序列预测的精度,本文利用希尔伯特-黄( Hilbert-Huang)变换理论中的经验模态分解( EMD)方法将非平稳时间序列分解,使之成为若干个频率单一的本征模态函数分量( IMF)。利用神经网络模型对各个本征模态函数进行预测,将各个预测结果进行重构加权,以提高预测精度。再结合历年旅客运输量来预测其后某段时间内运输量。实验结果表明,算法改进后的预测精度高于神经网络等预测方法。%In order to improve the prediction accuracy of non-stationary time series, the paper used Empirical Mode Decomposi-tion( EMD) method of Hilbert-Huang transform theory to decompose non-stationary time series into several IMF components of single frequency. Using the neural network model to predict IMF, the prediction results are reconstructed and weighted. The ac-curacy of prediction will be improved. It can also predict the transport volume in a certain period of time on the basis of the histor-ical passenger data. The experimental results show that the improved algorithm is better than the neural network method, etc.
展开▼