摘要
基于最小一乘准则对GM(1,1)模型解的任意常数项参数进行修正,通过引入统计量,给出估计表达式.数值结果表明,此方法与传统的GM(1,1)参数估计相比,稳健性更好.基于本方法对黑龙江省GDP建立GM(1,1)预测模型,并分析其影响因素,建立双对数模型,模型结果表明显著影响黑龙江GDP经济增长的为投资与R&D经费投入.最后为提高模型的预测精度,提出基于GM(1,1)模型和双对数模型的组合模型,经验证此模型预测相对误差较小.
We modify the parameters of arbitrary constant terms of the solution of the GM(1,1)model based on the least-absolute criterion.By introducing statistics,we write its expression Numerical results show that the method is more robust than traditional GM(1,1)parameter estimation.Using the above method,we establish a GM(1,1)forecasting model of Heilongjiang Province’s GDP.Additionally,we also analyze the main influencing factors of heilongjiang province’s GDP and establish a double logarithmic model.Consequentially,the main reasons for Heilongjiang’s GDP growth are investment and R&D expenditure.Finally,in order to improve the prediction accuracy of the model,we propose combination model of GM(1,1)model and the dual logarithm model As the result,the combination model has a smaller prediction relative error.
作者
周莹
曹连英
徐文科
ZHOU Ying;CAO Lian-ying;XU Wen-ke(Northeast Forestry University,Harbin 150040,China)
出处
《数学的实践与认识》
北大核心
2020年第23期267-274,共8页
Mathematics in Practice and Theory
基金
中央高校项目(2572018BC20)
黑龙江省自然科学基金(C201408)。