非等距GM(1,1)模型在沉降预测中的应用探讨

针对非等距沉降数据序列的建模问题，探讨了3种非等距GM(1,1)模型在沉降预测中的应用。结果显示，加权非等距GM(1,1)模型不适用于具有近似指数趋势的沉降数据，灰线性加权非等距GM(1,1)模型并非真正的非等距模型，模型预测式与时间无关，不能用于沉降预测，而对于笔者所建立的非等距GM(1,1)模型，其拟合函数等同于灰色线性回归组合模型，可用于近似指数趋势的沉降数据。通过实例分析对比了3种非等距GM(1,1)模型的应用效果，验证了上述观点的正确性。

Discussion of the Application of Non-Equidistant GM (1, 1) Models in Subsidence Prediction

According to the modeling problem of non-equidistant subsidence data sequence, the application of three kinds of non-equidistant GM (1, 1) models in subsidence prediction is discussed. The analysis results show that the weighted non-equidistant GM (1, 1) model is not appropriate for the subsidence data with approximate exponential trend, and the grey linear weighted non-equidistant GM (1, 1) model is not a really equidistant model. The prediction formula of the grey linear weighted non-equidistant GM (1, 1) model has nothing to do with the time, and cannot be used for subsidence prediction. The fitting function of the non-equidistant GM (1, 1) model established by the author is equivalent to that of grey linear regression combined model, and can be used on the subsidence data with approximate exponential trend. Finally, the effectiveness of the three kinds of non-equidistant GM (1, 1) model are compared by example analysis and this validates the correctness of the above viewpoints.