摘要
选取平顶山市2012—2018年GDP作为原始数据序列,建立传统GM(1,1)预测模型和背景值优化GM(1,1)预测模型.从级比偏差检验、后验差比检验和小概率误差检验可知,2个模型精度均为一级.背景值优化GM(1,1)模型的平均相对误差为0.0126,优于传统GM(1,1)模型.该模型对平顶山市GDP预测有很高的实用性.本文采用背景值优化GM(1,1)模型预测平顶山市2019—2023年的GDP.预测结果表明:平顶山市GDP 5 a内将保持平均增长率6.76%的速度平稳增长.
By selecting the GDP of Pingdingshan in 2012—2018 as the original data series,the prediction model of traditional GM(1,1)and background value optimization GM(1,1)are established.The accuracy of the two models is the first order according to the test of grade ratio deviation,posteriori difference ratio and small error probability.The average relative error of background value optimized GM(1,1)is 0.0126,which is better than the traditional GM(1,1).This model has high practicability to the GDP forecast of Pingdingshan.In this paper,the background value optimized GM(1,1)is used to predict the GDP of Pingdingshan in 2019—2023.The forecast result shows that the GDP of Pingdingshan will maintain an average growth rate of 6.76%in2019—2023.
作者
张晓果
赵颖范
杜亚冰
兰奇逊
李亚杰
ZHANG Xiao-guo;ZHAO Ying-fan;DU Ya-bing;LAN Qi-xun;LI Ya-jie(School of Mathematics&Physics,Henan University of Urban Construction,Pingdingshan 467036,China;Shenzhen Galaxy Holding Group Co.,Ltd.,Shenzhen 518046,China)
出处
《河南城建学院学报》
2022年第3期87-92,共6页
Journal of Henan University of Urban Construction
基金
国家自然科学基金项目(61503122)
河南省科技攻关计划项目(202102210142,212102210172)。
关键词
背景值优化
GM(1
1)
GDP
平均相对误差
后验差比
小误差概率
background value optimization
GM(1
1)
GDP
mean relative error
posteriori difference ratio
small error probability
