基于主分量的神经网络水位预报模型应用研究

Neural Network Prediction Model of Xijiang River Water Level Based on Principal Component Analysis

根据气象和水文资料,以上游面雨量、水位值为预报因子,以西江流域的梧州水位为预报量,发现预报因子与预报量有很好的相关性。采用人工神经网络与主分量分析相结合的方法,建立了梧州水位的预报模型。结果表明,该预报模型对历史样本拟合精度高,试报效果及预报稳定性明显好于传统的神经网络预报模型,可在预报业务中应用。

The analysis of the meteorological and hydrological data shows that there is close correlation between the water level of the Xijiang river and the upper reach water level and areal mean rainfall.The new neural network prediction model of water level is established based on the principal component analysis(PCA).The comparison between the new model based on PCA and the traditional neural network model indicates that the new model is significantly superior to the traditiona model in prediction accuracy and prediction stability,thus having a good prospect in operational application.