ARIMA模型对妊娠期糖尿病住院人数的预测

Prediction of the Number of Inpatients for Gestational Diabetes Mellitus with ARIMA Model

目的探讨自回归积分滑动平均模型(Autoregressive Integrated Moving Average Model,ARIMA)在妊娠期糖尿病住院病人预测中的应用,为妊娠期糖尿病(gestational diabetes mellitus,GDM)和母亲妊娠期糖尿病婴儿(infants of diabetes mothers,IDM)制定GDM和IDM健康管理工作提供依据。方法利用2013年1月1日-2017年12月31日湖北省妇幼保健院成人内科每月GDM住院患者数据,建立ARIMA模型预测GDM住院人数。结果通过参数和模型拟合优度检验以及残差白噪声序列检验,得到模型ARIMA(0,1,1)×(0,1,1)12,R2=0.439,BIC=3.999,Ljung-Box Q=15.714,P=0.473。预测2018年每月GDM住院人数实际值在预测值的95%可信区间内。结论ARIMA模型拟合效果较好,预测精度更高,可用于GDM住院人数的预测。

Objective To explore the application of ARIMA model to predict number of inpatients for gestational diabetes mellitus,so as to provide a scientific reference for more effective management of GDM and IDM.Methods The prediction model was built using the occurrence monthly number of inpatients for GDM from January 1 st 2013 to December 31 st 2019 in Hubei Provincial Maternal and Child Health Hospital and the predictive performance of this model was verified by the number of hospitalized patients of GDM in 2018.Results ARIMA(0,1,1)×(0,1,1)12 was built according to the parameter and model test of goodness of fit and Ljung-Box Q test(R2=0.439,BIC=3.999,Ljung-Box Q=15.714,P=0.473).The actual values of occurrence number of inpatients patients for GDM monthly in 2018 were among 95%CI.Conclusion ARIMA model has a good fitting effect and a high forecast precision,and can be used for predicting number of inpatients for GDM.