论文部分内容阅读
为优化神经网络模型的应用效果,研究了基于神经网络的降雨-径流模型,根据Copula熵法确定预报因子,并与传统的线性相关法进行比较分析,采用BP、RBF、GRNN三种神经网络建立降雨-径流模型,应用均方根误差、合格率、确定性系数三个指标为模型选取评价准则。通过对金沙江流域的径流预报,发现基于Copula熵法的BP模型预报结果更接近实测值,精度更高。
In order to optimize the application effect of neural network model, the rainfall-runoff model based on neural network is studied. The forecasting factor is determined by Copula entropy method and compared with the traditional linear correlation method. BP neural network, RBF neural network and GRNN neural network are established Rainfall-runoff model, root mean square error of application, pass rate and deterministic coefficient were selected as evaluation criteria. By forecasting the runoff in the Jinsha River, it is found that the BP model predictions based on the Copula entropy method are closer to the measured values and have higher accuracy.