基于ARIMA和GM(1,1)模型的我国农村MMR、IMR、NMR及U5MR趋势预测  

Prediction of trends for MMR,IMR,NMR,and U5MR in rural China based on ARIMA and GM(1,1)models

作　　者：张彬 张龙秀 王瑞 曹武军 刘辉 ZHANG Bin;ZHANG Longxiu;WANG Rui;CAO Wujun;LIU Hui(Department of Science and Education,Anhui Province Maternity&Child Health Hospital,Anhui,Hefei 230001,China;Imaging Center,Anhui Provincial Children’s Hospital,Anhui,Hefei 230041,China;Hefei Municipal Health Commission,Anhui,Hefei 230071,China)

机构地区：[1]安徽省妇幼保健院科教处,安徽合肥230001 [2]安徽省儿童医院影像中心,安徽合肥230041 [3]合肥市卫生健康委员会,安徽合肥230071

出　　处：《中国医药科学》2023年第14期7-10,15,共5页China Medicine And Pharmacy

基　　金：中国疾病预防控制中心妇幼保健中心“母婴营养与健康研究项目”(2022FYH007);安徽省社会科学创新发展研究课题(2021CX520);安徽省教育厅科学研究项目(2022AH050630);安徽卫生思想政治工作促进会(CJH2022005)。

摘　　要：目的分析我国农村孕产妇、婴儿、新生儿及五岁以下儿童四种死亡率指标变化趋势并进行预测,为卫生行政部门制定农村母婴安全发展战略提供参考依据。方法采用ARIMA和GM(1,1)模型分别预测2021—2025年四个指标变化趋势。结果ARIMA(0,1,0)模型和GM(1,1)模型对四个指标预测的平均绝对误差分别为0.908、0.410、0.194、0.584和0.718、0.297、0.128、0.341,平均相对误差分别为0.041、0.040、0.030、0.043和0.032、0.029、0.020、0.025,GM(1,1)预测误差较小。结论运用GM(1,1)模型对四个指标进行拟合和预测效果优于ARIMA模型。Objective To analyze the trend of changes in four mortality indicators of pregnant women,infants,newborns,and children under five years old in rural areas of China,and to predict them,in order to provide a reference for the Health Administrative Department to formulate strategies for the safe development of rural mothers and infants.Methods The ARIMA and GM(1,1)models were used to predict the trends of four indicators from 2021 to 2025.Results The mean absolute errors of ARIMA(0,1,0)model and GM(1,1)model for the prediction of the four indicators were 0.908,0.410,0.194,0.584 and 0.718,0.297,0.128,0.341,respectively,and the mean relative errors were 0.041,0.040,0.030,0.043 and 0.032,0.029,0.020,0.025,respectively.The prediction error of GM(1,1)was small.Conclusion The GM(1,1)model is superior to the ARIMA model in fitting and predicting the four indicators.

关 键 词：时间序列模型 灰色模型 孕产妇死亡率 婴儿死亡率 新生儿死亡率 五岁以下儿童死亡率 

分 类 号：R17[医药卫生—妇幼卫生保健]