WRBF 网络模型在月径流量预测中的应用

摘要

为适应径流序列的不确定性及随机性对预测的影响，将小波变换和 RBF 网络相结合构造了 WRBF 网络模型。该模型综合了小波分析的多分辨率优势和 RBF 网络的非线性逼近功能。黑河流域的模拟结果表明：WRBF 网络的稳定性较好，预测合格率较高。%In order to adapt the influence of the uncertainty and randomness of runoff sequence on its forecast,the wavelet transform was combined with RBF neural network to establish the WRBF network model. The model integrated the multi-resolution advantages of the wavelet analysis and the nonlinear approximation function of the RBF network. The simulation results of the Heihe River basin show that the stability of the WRBF model is better and its qualified rate of prediction is higher.