Flood forecasting method of coupling wavelet transform with BP neural network
Zhijia Li · Journal of Hydroelectric Engineering · 2009
In order to explore the value of coupled wavelet transform with BP neural network in flood forecasting,a flood forecasting method is proposed based on the coupling.Two structures of the model are designed according to the different coupled mode.They are the loose coupled model and compact coupled model.The two models are trained and verified with measured runoff time series respectively.The calculated results indicate that the forecast effects of the two models are improved with different degrees compared with three-layer normal BP neural network,the forecast effect of the loose coupled model is generally better,therefore,coupled wavelet transform with BP neural network has better opplication prospect in flood forecasting.