基于灰色GM（1，1）模型的城市生活用水量预测

介绍灰色理论建模原理和模型参数辨识方法，并以上海市为实例建立灰色GM（1，1）预测模型，运用残差检验、后验差检验以及关联度检验3种方法对模型进行精度检验，其模型拟合精度达97．25％。用所建立的模型对上海市2003～2007年城市生活用水量进行预测。结果表明，该灰色模型用于城市生活用水量预测，符合其灰色特性，通用性好，并且所需数据少，计算量适中，预测结果与实际情况比较吻合。

Prediction of city living water requirement based on GM(1,1) model

The modeling principle of grey theory and the recognition method of model parameter were presented. Taking Shanghai City as a case, a grey GM（ 1,1 ） model was established based on the data of city water requirement from 1988 to 2002. The precision test of residual error examination, posterior difference examination and relative degree examination showed that the fitting precision reached 97.25 %. The model was applied to the prediction of city living water requirement in Shanghai city from 2003 to 2007. The result indicates that this grey model used in city living water requirement prediction conforms to its grey characteristic. The model has the advantages of good versatility, few requirements of data, moderate computation quantity, and the result accords with the practical condition.