不同径流尺度的小波神经网络预测

Forecasting Approach for Different Scales of Runoff Based on Wavelet Analysis and Neural Network

探讨了小波神经网络在不同径流时间序列尺度预测中的应用,结合神经网络和小波转化的优点,弥补了人工神经网络在预测应用中的不足.并将小波神经网络预测结果与人工神经网络预测结果进行比较,结果表明:小波神经网络的预测结果与实测值之间的误差明显小于人工神经网络的预测值,显示出小波神经网络具有较高的预测精度.

This article discussed the application of the wavelet analysis and neural network method in runoff prediction. This method unified the merits of the neural networks and the wavelet transformation. The wavelet analysis and neural network method made up the defects of artifical neural network in forecast application. Comparison was made between the wavelet analysis and neural network method and the artificial neural network. The results indicated that the wavelet analysis and neural network method has high accuracy for the runoff prediction.