非等间隔灰色线性回归组合模型估参方法改进及应用

Improvement and Application of Parameter Estimation Method of Unequal Interval Grey Linear Regression Combination Model

针对灰色线性回归组合模型估参较传统灰色GM(1,1)模型估参复杂,且不易掌握和使用的问题,构建了该类模型所满足的白化微分方程,提出了一种与传统灰色模型思路相同的估参方法,该方法高效实用,且易掌握.另外,在等间隔灰色线性回归组合模型的基础上建立了某楼宇沉降预测的非等间隔灰色线性回归组合模型,并用改进后的估参方法计算相关参数.实证分析表明所建模型具有较高的拟合预测精度,可用以指导工程实践.

Aiming at the problem that the parameter estimation of grey linear regression combined model is more complex than that of traditional grey GM(1,1)model,and it is not easy to master and use,the whitening differential equation satisfied by this kind of model is constructed,a parameter estimation method with the same idea as the traditional grey model is proposed,which is efficient,practical and easy to master.In addition,based on the equal-interval grey linear regression combination model,a non-equal-interval grey linear regression combination model for the settlement prediction of a building is established,and the relevant parameters are calculated by the improved parameter estimation method.The empirical analysis shows that the model has highfitting and prediction accuracy,which can be used to guide engineering practice.