改进背景值的非等间距GM(1,1)模型预测用水量

Application of Non-equidistance GM(1,1) Model with Improved Background Value in Predicting Water Demands

考虑到在工程实际中用水资料可能有所缺失或因某些数据异常而造成数据列的不等间距,构造了非等间距GM(1,1)模型,并对其背景值进行改进。结果表明,改进背景值后非等间距GM(1,1)模型的预测精度大大提高,用其预测城市用水量是合理可行的。

Considering the possible missing in water demand data and the non-equidistance of data series caused by data abnormity, the non-equidistance GM （1,1） model was established, and its back- ground value was improved. The results show that the prediction accuracy of non-equidistanee GM（1,1） model with improved background values is greatly improved, and it is reasonable and feasible to predict the urban water demands by the model.