SARIMA模型在河南省急性出血性结膜炎发病预测中的应用

Application of SARIMA model in forecasting the incidence of acute hemorrhagic conjunctivitis in Henan Province

目的:应用SARIMA模型预测河南省急性出血性结膜炎发病人数。方法:使用河南省2013年1月至2020年6月急性出血性结膜炎月发病数建立SARIMA模型,利用2020年7至12月的逐月发病数据验证SARIMA模型的预测效果,并进一步预测2021年1至12月河南省急性出血性结膜炎发病人数。结果:急性出血性结膜炎具有明显的季节性周期特点,每年的6至8月为发病的高峰期,季节性周期约为1 a。建立预测模型SARIMA(0,1,1)(0,1,1)12,其表达式为ΔΔ12lnYt=(1-0.730L)(1-0.671L12)εt,Ljung-Box检验统计量为0.902,残差序列为白噪声序列(P=0.342),AIC=-48.183,BIC=-41.753。2020年7至12月预测值与实际值相比,平均绝对误差为24.50,平均相对误差为10.28%。预测2021年6月发病人数达到当年高峰(281人),整体的发病趋势与2020年相比略有下降。结论:SARIMA(0,1,1)(0,1,1)12模型可用于河南省急性出血性结膜炎发病人数的短期预测。

Aim:To apply the SARIMA model in forecasting the incidence of acute hemorrhagic conjunctivitis in Henan Province.Method:SARIMA model was established based on the monthly reported cases data of acute hemorrhagic conjunctivitis in Henan Province from January 2013 to June 2020.The monthly incidence data from July 2020 to December 2020 were used to verify the prediction effect of the SARIMA model.And then this model was used to invert the number of acute hemorrhagic conjunctivitis in Henan Province from January 2021 to December 2021.Results:Acute hemorrhagic conjunctivitis had obvious seasonal cycle characteristics,and the peak of the disease occurred from June to August every year,with a seasonal cycle of about one year.A model of SARIMA(0,1,1)(0,1,1)12 was established,with the equation asΔΔ12 lnYt=(1-0.730L)(1-0.671L12)εt,the Ljung-Box test statistic value was 0.902,residual error sequence was white noise sequence(P=0.342),AIC=-48.183,BIC=-41.753.The average absolute error was 24.50 and the average relative error was 10.28%,respectively when compared with the simulated number of patients from July 2020 to December 2020.It was predicted that the number of cases in 2021 would reach the peak in June(281),and the overall incidence trend would decrease slightly compared with 2020.Conclusion:The model SARIMA(0,1,1)(0,1,1)12 could be used to predict the incidence of acute hemorrhagic conjunctivitis in Henan Province.