基于广义回归神经网络(GRNN)的地下水水位预测

The prediction of the groundwater table based on GRNN

传统的地下水位预测方法存在需要事先假定含水层是均匀规则的且计算量大等不足。基于广义回归神经网络基本原理,建立了地下水位预测的广义回归神经网络模型。通过实例计算表明,该模型在信息有限的情况下,仍可达到满意的预测精度,且模型简单,计算效率较高,具有广阔的应用前景。

The traditional method of underground water level prediction has lots of deficiencies.For example ,it needs to be assumed that the aquifer is homogeneous and well-regulated ,and the calculation amount is quite large as well.Based on the basic principle of GRNN ,the GRNN mode of groundwater table prediction is built. With examples and calculations ,it shows that this mode can still achieve satisfactory prediction accuracy ,even in the case of information limitation.Moreover, the mode is simple, it also has relatively high computational ef- ficiency and broad prospect of application.