Runoff Forecast by Combination Wavelet Transform with the Back Propagation Artificial Neural Networks

摘要

In this paper, the runoff sequence has been predicted with two kinds of methods separately. One is the Back Propagation Artificial Neural Networks (BPANN), the other is the combination wavelet transformation with BPANN. And some comparison with the result that is predicted has done. We draw the following conclusions: if there are no rainfall and other information, predicting the runoff sequence only with runoff material is difficult for us to make the satisfied result. The method of the combination wavelet transformation with BPANN can make better result in forecasting of the runoff sequence. The course of the runoff sequence predicted is introduced in detail. At the same time, the factors influencing the numerical results are investigated and useful suggestions are presented for other applications of the wavelet transform to runoff analysis.