应用GM(1,1)模型预测2015年我国九市郊区儿童初生体格发育指标缺失值

Using GM(1,1)to predict the missing values of primary physical development of children in nine suburban areas of China in 2015

目的预测2015年中国九市郊区儿童初生体格发育指标缺失值。方法以1975—2005年的4次中国九市城郊儿童体格发育调查收集的郊区儿童初生体质量、身长、头围数据分别建立原始序列,或引入平均弱化缓冲算子,运用均值GM(1,1)预测2015年初生体格发育指标的缺失值。通过绝对关联度(σ)、平均相对误差(α)、均方差比值(C)、小误差概率(P)判断预测模型的精度等级。结果郊区男童初生体质量GM(1,1)发展系数a=-0.0153、灰作用量b=3.1463,σ=0.9997、α=0.0045%、C=0.0040、P=1.00,2015年预测值为3.3712 kg;女童初生体质量引入2阶弱化缓冲算子,GM(1,1)发展系数a=-0.0017、灰作用量b=3.1724,σ=0.9995、α=0.0391%、C=0.1850、P=1.00,2015年预测值为3.1962 kg。男童初生身长GM(1,1)发展系数a=-0.0020、灰作用量b=50.0503,σ=0.9999、α=0.0001%、C=0.0005、P=1.00,2015年预测值为50.5003 cm;女童初生身长GM(1,1)发展系数a=-0.0020、灰作用量b=49.4501,σ=0.9999、α=0.0001%、C=0.0006、P=1.00,2015年预测值为49.9003 cm。男童初生头围引入2阶平均弱化缓冲算子,GM(1,1)发展系数a=-0.0003、灰作用量b=34.1649,σ=0.9984、α=0.0144%、C=0.2994、P=1.00,2015年预测值为34.2148 cm;女童初生头围引入4阶平均弱化缓冲算子,GM(1,1)发展系数a=0.0003、灰作用量b=33.7320,σ=0.9993、α=0.0049%、C=0.1981、P=1.00,2015年预测值为33.6924 cm。结论均值GM(1,1)预测模型精度等级高,预测结果较可靠。

Objective To predict the missing values of primary physical development of children in nine suburban areas of China in 2015.Methods The original sequence based on birth body mass,length and head circumference of suburban children from 1975 to 2005 was established in nine cities of China,or introduced the average weakening buffer operator,and the missing values of primary physical development in 2015 were predicted by the mean GM(1,1)model.The accuracy level of the models were judged by absolute degree of incidence(σ),mean relative error(α),mean square deviation ratio(C),and small error probability(P).Results The development coefficient of GM(1,1)for the birth body mass of rural boys was a=-0.0153,grey action quantity was b=3.1463,σ=0.9997,α=0.0045%,C=0.0040,P=1.00,and the predicted value of boy′s birth body mass in 2015 was 3.3712 kg;the girl′s model introduced the 2 order weakening buffer operators,the development coefficient of GM(1,1)for the birth body mass of rural girls was a=-0.0017,grey action quantity was b=3.1724,σ=0.9995,α=0.0391%,C=0.1850,P=1.00,and the predicted value of girl′s birth body mass in 2015 was 3.1962 kg.The development coefficient of GM(1,1)for the birth length of rural boys was a=-0.0020,grey action quantity was b=50.0503,σ=0.9999,α=0.0001%,C=0.0005,P=1.00,and the predicted value of boy′s birth length in 2015 was 50.5003 cm;the development coefficient of GM(1,1)for the birth length of rural girls was a=-0.0020,grey action基金项目:江苏省妇幼健康科研项目(F2018035)通信作者:李鸿斌,Email:lihb@163.comquantity was b=49.4501,σ=0.9999,α=0.0001%,C=0.0006,P=1.00,and the predicted value of girl′s birth length in 2015 was 49.9003 cm.The boy′s model introduced the 2 order weakening buffer operators,the development coefficient of GM(1,1)for the birth head circumference of rural boys was a=-0.0003,grey action quantity was b=34.1649,σ=0.9984,α=0.0144%,C=0.2994,P=1.00,and the predicted value of boy′s birth Head circumference in 2015 was 34.2148cm;the girl′s model introduced the 4 order weakening buffer operators,the development coefficient of GM(1,1)for the birth head circumference of rural girls was a=0.0003,grey action quantity was b=33.7320,σ=0.9993,α=0.0049%,C=0.1981,P=1.00,and the predicted value of girl′s birth head circumference in 2015 was 33.6924 cm.Conclusion The Mean GM(1,1)has high accuracy and reliable prediction results.