基于区间序列的GM(1.1)地下水水位预测

为探讨基于区间序列的GM（1.1）模型预测地下水水位的准确性和信息量，本文选取界点建模、转换建模和参数建模3种区间方法进行比较和分析，结果表明：界点的相对误差均值都小于0.01%，区间相对误差均值小于7%；区间序列的GM（1.1）模型具有一定的延迟效应和紊乱效应；实际水位值波动幅度较小时模型预测精度较高；参数建模的预测精度较优于界点建模和转换建模。与点预测相比，区间序列的GM（1.1）模型预测的地下水水位变化规律更有利于地下水资源管理与开发利用。

PREDICTION OF GROUNDWATER LEVEL BASED ON INTERVAL SERIES GM(1. 1)MODEL

To investigate the accuracy and information of GM(1. 1)model bases on interval series in predicting groundwater level in this paper, three interval methods of boundary point modeling, transformation modeling and parameter modeling are selected for comparison and analysis. The GM(1. 1)model of interval sequence has some delay effect and disorder effect. When the fluctuation range of actual water level is small, the prediction accuracy of the model is higher. The prediction accuracy of parameter modeling is better than that of boundary point modeling and transformation modeling. Compared with the point prediction, the interval series GM(1. 1)model is more beneficial to groundwater resource management and development.