摘要
为提高非线性和非平稳性的棉花期货价格的预测能力,该文采用基于EMD的BP神经网络预测方法,对原始序列进行分解,产生多个平稳序列imf,然后对每个imf进行预测,把预测结果相加得到原序列的预测值。实验结果表明,基于EMD的BP神经网络预测方法优于单纯BP网络预测。
To improve tile predictive ability of nonlinear and non-stationary cotton futures price,we adopt the BP neural network prediction method based on EMD. The method first decomposes the original sequence,thus get- ting a few stationary series, and then the study predicts every imf, finally adding all the predicted values above to get the predicted value of the original sequence. The result of research shows that the BP neural network predica- tion method based on EMD works better than the simple BP network prediction method.
出处
《石河子大学学报（哲学社会科学版）》
2013年第1期78-80,共3页
Journal of Shihezi University(Philosophy and Social Sciences)
