摘要
目的:探讨自回归移动平均乘积季节模型(SARIMA)在成都市龙泉驿区出生缺陷率预测中的构建与应用的可能性。方法:将出生缺陷监测数据通过预处理、模型识别、参数估计和模型诊断建立SARIMA模型,并对模型进行评价与预测。结果:成都市龙泉驿区的出生缺陷发生率为一个非平稳时间序列且具有一定的季节性。通过自相关系数和偏相关系数建立了ARIMA(1,1,2)×(1,1,1)12乘积季节模型,该模型的残差经白噪声检验(Ljung-Box检验)为白噪声序列(P>0.05);在预测2019年10月至2019年12月的出生缺陷发生率时,平均相对误差仅为3.89%,并预测2020年龙泉驿区的平均出生缺陷发生率为178.41/万。结论:ARIMA(1,1,2)×(1,1,1)12乘积季节模型适用并可较为精确地短期预测成都市龙泉驿区的出生缺陷发生率。
Objective:To explore the possibility of seasonal autoregressive integrated moving average(SARIMA)model in the prediction of birth defects incidence rate in Longquanyi District,Chengdu.Methods:Through preprocessing,model recognition,parameter estimation and model diagnosis,the birth defects monitoring data are used to build SARIMA model,and the model is evaluated and predicted.Results:The incidence rate of birth defects in Longquanyi District of Chengdu is a non-stationary time series with a certain seasonality.ARIMA(1,1,2)×(1,1,1)12 model was established through autocorrelation coefficient and partial correlation coefficient,and the residual error of the model was white-noise sequence tested by Ljung-box test(P>0.05).In addition,the model predicting the incidence rate of birth defects from October to December 2019 that the average relative error is only 3.89%,and the average incidence rate of Longquanyi District of birth defects in 2020 is 178.41 per 10000.Conclusion:The ARIMA(1,1,2)×(1,1,1)12 model is suitable for predicting the incidence rate of birth defects in Longquanyi District of Chengdu in a short term.
作者
徐寰宇
熊利英
郭玲
张雪梅
XU Huanyu;XIONG Liying;GUO Ling;ZHANG Xuemei(Primary Health Care Department of Longquanyi District Maternal and Child Health Hospital,Chengdu 610100,China)
出处
《现代医学》
2021年第10期1144-1148,共5页
Modern Medical Journal
基金
四川省妇幼保健协会科研项目(2020YB12)。
关键词
出生缺陷
围产儿
自回归移动平均乘积季节模型
预测
birth defects
perinatal infant
seasnoal autoregressive integrated moving model
prediction
