统计与决策2026，Vol.42Issue(15)：51-57,7.DOI:10.13546/j.cnki.tjyjc.2026.15.008
含时间多项式项的单变量灰色模型识别方法及其应用
Identification Method for Univariate Grey Model With Time Polynomial Terms and Its Application
摘要
Abstract
To address the problem that the truncation parameter is difficult to determine in univariate grey models with time polynomial terms,this paper proposes a criterion for determining the truncation parameter based on the difference idea from the characteristics of the original data.The paper first analyzes the expression forms of typical grey models with time polynomial terms to derive the general formula of the time response function of such models,then uses the difference idea to establish the quantita-tive relationship between the truncation parameter and the intrinsic law of the original data,thereby obtaining a method obtained for determining the truncation parameter without human intervention,and finally,applies the proposed method to the prediction of China's total nuclear energy consumption.The results show that the model determined by this method has high accuracy in the fit-ting stage and exhibits good extrapolation performance in the prediction stage,which is significantly better than the traditional em-pirical selection method.The research results also show that this criterion can adaptively determine the truncation parameter ac-cording to the characteristics of the original data,effectively solving the problem of model structure identification for grey models with time polynomial terms,thus providing a reliable technical approach for the practical application of such models in grey pre-diction theory.关键词
时间多项式项/灰色模型/差分思想/模型识别Key words
time polynomial term/grey model/difference idea/model identification分类
自科综合引用本文复制引用
樊学秀,覃素平,许泽东..含时间多项式项的单变量灰色模型识别方法及其应用[J].统计与决策,2026,42(15):51-57,7.基金项目
国家自然科学基金资助项目(72271120 （）
72001093) （）
