加权组合模型在城市用水量预测中的应用

Application of Weighted Composition Model in Urban Water Consumption Forecasting

鉴于传统灰色模型在建模中存在固有偏差的问题,本文采用无偏灰色GM(1,1)预测模型.在无偏灰色预测模型表达式的基础上,又提出了非线性的预测模型,并将其用于城市的用水量预测上.考虑到单一预测模型在预测过程中存在的不足,用最优加权组合模型对无偏灰色GM(1,1)模型和非线性模型进行组合,并将加权组合模型首次用于遂宁市的城市用水量预测.预测结果表明,组合模型的预测精度优于单一的预测模型,预测结果与城市的实际用水量拟合较好,该方法可推广到其它类似城市的用水量预测中.

In view of the inherent deviation existing in the traditional grey model during modeling,this article used unbiased grey GM（1,1） forecasting model.Based on the expression of the unbiased grey forecasting model,the nonlinear forecasting model was put forward and applied for forecasting urban water consumption.Considering the deficiency of the single forecasting model in the course of forecast,the unbiased grey GM（1,1） forecasting model was combined with nonlinear model by weighted composition mode,which was firstly applied for forecasting urban water consumption of Suining city,Sichuan province.The prediction results indicated that the combination model was superior than the single model in predictive precision and the results show that this method provides a better fit to the real urban water consumption,which can be popularized to the water consumption prediction of other similar cities.