基于GM(1,1)模型的沉陷观测数据处理

Subsidence data processing based on the observation GM(1,1) model

使用非等间距GM(1,1)模型和基于插值的GM(1,1)模型对某开采面的下沉情况进行了预计,并对不同大小插值的GM(1,1)模型的精度进行分析,由预计结果比较了两类模型的优缺点.两种模型的预测结果的残差都较大.因此,使用灰色模型对矿区的某些点的变化情况进行预测时,应该注意预测的时间段不应太长,对于较短时间间隔的沉陷预计,上述模型能够得到较高精度的预计结果,可以为矿区的沉陷趋势的分析和研究提供一定的依据.

Using the non-equidistant GM （1,1） model and interpolation based GM （1,1） model, a mining surface subsidence case was expected. The accuracy of different sizes interpolation GM（1, 1） model was analyzed and the advantages and disadvantages of the two models were compared by the expected results. It was found that the predictions of both two models had large residuals. So, when using grey models to predict the changes of some mining spots, the time intervals of the prediction should not be too long. The models can get accurate prediction results for mining set- tlement estimate with short time intervals, which could have some guidance for mining subsidence research.