摘要
文章以嫩江大赉站1970-2012年的年蒸发量为例,首先根据AIC准则来确定模型的阶数,进而确定RBF神经网络的输入向量,建立了RBF网络年蒸发量预报模型。结果表明,RBF神经网络模型能很好的预报年蒸发量,同BP网络比,RBF网络预测的稳定性更好,训练速度更快,预测精度更高。RBF年蒸发量预报模型可以有效地弥补物理模型的不足,而且对于不同地区具有普适性。
Taking annual evaporation of Dalai station of Nenjiang River from 1970 to 2013 as an example, the paper determines the order of model with AIC norm and the input vector of RBF neural network, establishes annual evaporation forecasting model based on RBF network. The results show the RBF neural network model could forecast annual evaporation very well, its stability, training speed and accuracy are better than BP neural network. The model could make up for the lack of physical model and has the universal applicability of the different regions.
出处
《东北水利水电》
2014年第11期36-37,72,共2页
Water Resources & Hydropower of Northeast China
关键词
RBF神经网络
年蒸发量
预报
嫩江大赉站
RBF neural network
annual evaporation
forecasting
Dalai station of Nenjiang River