摘要
为改进灰色GM(1,1)组合模型在装备维修器材消耗需求预测中随机数据的拟合效果,提高预测精度,提出将灰色GM(1,1)模型和多项式回归模型相结合的方法,深入挖掘序列中除单一指数函数规律外的复杂函数变化规律,建立灰色多项式回归模型;模型首先通过将函数展开为Taylor级数,讨论了运用灰色多项式回归组合模型表征除单一指数规律外的复杂函数形式的合理性;随后给出了灰色多项式回归组合模型指数项系数的求解过程;最后通过运用最小二乘法得到灰色多项式回归组合模型预测式的各项系数;分别运用灰色2项、3项、4项式回归组合模型以及灰色GM(1,1)模型针对同一组历史数据进行了计算,结果表明:灰色多项式回归组合模型可大幅度提高数据的拟合精度,并且随着拟合项数的增加,拟合精度逐渐提高。
In order to improve the fitting effect and precision of GM(1,1)model in the prediction of consumption demand of equipment maintenance materials,the GM(1,1)model and polynomial regression model were combined to study the complex law,besides the single exponential law.The Grey polynomial regression combined model was constructed.Taylor series was used to study the rationality of grey polynomial regression combined model,and the coefficient of exponential function part was solved.Finally,the coefficients of Grey polynomial regression combined model were solved by least square method.Grey binomial,trinomial and quadrinomial combined models and GM(1,1)model are used to study the same set of data.The result shows that,Grey polynomial regression combined model can improve the precision effectively,and the more terms of Grey polynomial use,the more precise the regression model is.
作者
张磊
于战果
李世民
ZHANG Lei;YU Zhanguo;LI Shimin(Army Academy of Border and Coastal Defense,Xi'an 710108,China;Army Military Transportation University,Tianjin 300161,China;The No.63963 rd Troop of PLA,Beijing 100072,China)
出处
《兵器装备工程学报》
CAS
北大核心
2019年第1期179-183,共5页
Journal of Ordnance Equipment Engineering
基金
全军后勤科研计划项目
关键词
灰色多项式回归模型
灰色预测模型
复杂函数
预测精度
Grey polynomial regression model
Grey prediction model
complex model
prediction precision
