【摘要】 基于BP神经网络,建立了一种综合时间序列分析和多元分析特点的动态水量预测模型,模型除了将影响用水量的因素作为输入节点之外,还将预报日前2 d的用水量作为输入节点,使得模型不但反映了用水量与影响因素的关系,还揭示了用水量时间序列的非线性特性.经生产实践检验,该模型的预测精度达到工程要求.更多还原

【Abstract】 Based on BP model of the artificial neural network, this paper establishes a dynamic water consumption-forecasting model with the characteristic of both time series and regression. This model use the factor affecting water consumption and water consumption two days forward as input node, so it not only reflects the relation between water consumption and its factors, but also shows the nonlinear characteristic of time series. According to practice in our production, the forecasting precise of this model fits the engineering demand completely.更多还原