一种状态区间取值修正的灰色-马尔科夫链预测模型研究

Research on a Gray-Markov Chain Prediction Model Based on the Range Correction of State Intervals

提出了一种对灰色-马尔科夫链预测模型的状态区间取值修正方法。考虑实测值在状态区间中的位置分布情况作为先验知识,对状态区间的端点重新定权,修正预测值在状态区间中的取值,从而获得更高精度的预测结果。结合某建筑物的实测沉降数据,与原始GM(1,1)、残差修正GM(1,1)、灰色-马尔科夫链预测模型进行对比,结果表明:本文预测模型精度更高,状态区间取值修正方法对灰色-马尔科夫链的精度改善有一定的实用性。

This paper presents a method based on the range correction of state intervals for the gray-Markov chain prediction model.Considering the positional distribution of measured values in state intervals as a priori knowledge,the endpoints of state intervals were re-weighted to modify the predictive values in state intervals,for which the more accurate forecasting results would be given.Combined with the measured data of a building settlement and compared with the original GM(1,1),GM(1,1)with modification of residuals,and gray-Markov chain prediction model,the results show that the prediction model in this paper has higher precision,and the range correction method of state intervals has certein applicability to improve the accuracy of the gray-Markov chain.