摘要
建立指数平滑和灰色模型组合预测模型,采用改进熵值法确定权重系数,并运用组合模型对宁波市进行了需水预测。以宁波市2001~2015年城市用水总量为基础数据,2016年城市用水总量作为测试数据,预测结果显示,组合预测模型的相对误差只有0.17%,低于指数平滑预测的1.71%和灰色模型预测的1.19%,组合预测模型进一步提高了单一预测模型的预测精度。
The combination forecasting model of exponential smoothing and grey model was established.Improved entropy method was used to determine the weight coefficient, and the combination model was used to predict the water demand of Ningbo City. Based on the total urban water consumption in Ningbo from 2001 to 2015, the total urban water consumption in 2016 was taken as the test data. The prediction results showed that the relative error of the combination forecasting model was only 0.17%, which was lower than 1.71% of the exponential smoothing forecasting model and 1.19% of the grey model forecasting. The combination forecasting model further improved the prediction accuracy of the single forecasting model.
引文
[1]水科学进展编辑部.需水管理的定义与内涵[J].水科学进展,2009,20(5):747-754.
[2]陈华友.组合预测方法有效性理论及其应用[M].北京:科学出版社,2008.
[3]李振全,徐建新,邹向涛,等.灰色系统理论在农业需水量预测中的应用[J].中国农村水利水电,2005,11:24-26.
[4]陈华友.熵值法及其在确定组合预测权系数中的应用[J].安徽大学学报(自然科学版),2003,27(4):1-6.
[5]宁波市水利局.宁波市水资源公报[R].2001-2016.