无偏灰色—马尔科夫组合模型在变形预测中的应用

Unbiased GM-Markov model in deformation prediction

在传统无偏灰色预测模型和马尔科夫模型的基础上提出了一种新的组合模型来对变形监测中的沉降进行预测。首先针对传统灰色预测模型存在灰色偏差与抗干扰能力弱的局限性,以及随机波动大的动态过程,提出采用无偏灰色—马尔科夫组合模型进行沉降预测。实验结果表明,利用该组合方法预测结果的精度要明显优于无偏灰色预测。

This paper proposes a new combination model to forecast the settlement deformation on the basis of the traditional unbiased GM prediction model and Markov model.In view of the traditional grey prediction model existing gray deviation and the limitation of the weak anti-interference ability,and the dynamic process of stochastic volatility,unbiased GM-Markov model is proposed for settlement prediction.Experimental results show that using the method of the combination,the precision of prediction results is improved significantly,and it is superior to the unbiased GM method.