中文摘要：

为了更加准确地预测国控监测点的日取用水量,对取用水量预测模型进行了研究。由于国控监测点的取用水量受多种因素的影响,数据具有非线性、非平稳的特性,应用传统的模型预测误差较大。为解决上述问题,提出了右延拓-镜像延拓EMD-BNN（REME-EMD-BNN）的预测模型,并以华北地区水厂A的日取用水量为例进行仿真分析,与传统的贝叶斯神经网络（BNN）、镜像延拓EMD-BNN、支持向量机（SVM）、镜像延拓EMD-SVM模型进行比较,结果显示,新提出的模型比传统模型的预测精度高。证明REME-EMD-BNN模型能更加准确预测国控监测点的日取用水量。

英文摘要：

In order to enhance prediction precision of the water intake quantity of national control monitoring points,a prediction model of water intake quantity is studied. The water intake quantity of national control monitoring points can be affected by many factors and exhibit nonlinear and non- stationary. The prediction error of traditional model is larger. For this reason,a new prediction model right extension mirror extension EMD- SVM（ REME- EMD- BNN） is researched,the model is used to predict water intake quantity of waterworks A in North China,and compared with the traditional BNN,mirror extension EMD- BNN,SVM,and mirror extension EMD- SVM models. The results demonstrate that the prediction accuracy of the new model is higher than the other traditional models. Therefore,the REME- EMD- BNN prediction model is able to efficiently improve the prediction accuracy of the water intake quantity of national control monitoring points.