变权组合预测模型在地下水水位预测中的应用

Application of Changeable Weight Combination Forecasting Model to Groundwater Level Prediction

为探讨变权组合预测模型在地下水水位预测中的应用,以内蒙古通辽市东郊为研究区,利用该区1997—2008年的地下水水位埋深数据,采用基于指数预测法、灰色预测法以及线性回归预测法的变权组合预测方法进行水位埋深的模拟和预测。经计算,以上3种单项预测方法的预测值以及组合预测方法的预测值与实际值的误差平方和分别为5.1161、5.0801、4.9148、4.6723。这表明,利用变权组合预测方法比单纯运用某一种预测方法预测精度更高。并且,由预测值可以看出,到2015年该研究区水位埋深将达到15.39 m,相比2008年递增幅度将达到20%。

In order to discuss the application of changeable weight combination forecasting model in groundwater levels forecasting, based on the date of groundwater levels from 1997 to 2008 in eastern suburbs of Tongliao City, Inner Mongolia Autonomous Region, the changeable weight combination forecasting method which is based on index model, grey model and linear regression model was used to simulate and forecast groundwater levels of the research area. It was calculated that the sum squares of errors of various forecasting were 5. 116 1,5. 080 1,4. 914 8,4. 672 3 respectively. The results indicated that the changeable weight combination forecasting method was more precise than one single forecasting method. From the predictive value, the depth of groundwater level in this research area will decrease to 15.39 m by the end of 2015, and compared with the depth of groundwater level in 2008, the decreasing range will come up to 20%.