上城区其他感染性腹泻ARIMA乘积季节模型的建立与预测

目的建立上城区其他感染性腹泻病求和自回归移动平均(auto-regressive integrated moving average,ARIMA)乘积季节模型,为早期防控提供参考。方法利用SPSS 25.0软件对上城区2010—2020年其他感染性腹泻病发病数据构建ARIMA乘积季节模型,通过对2021年月发病数进行回代预测评价模型拟合效果,并用构建的模型对2022年月发病数进行预测。结果上城区2010—2020年共报告其他感染性腹泻病40534例,年均报告发病数为3685例,无死亡病例报告。构建的较优模型为ARIMA(2,1,1)(1,1,1)12,平稳R2=0.870,贝叶斯信息准则(Bayesian information criterion,BIC)=9.524,平均绝对百分误差(mean absolute percentage error,MAPE)=27.351,模型Box-Ljung检验差异无统计学意义(Q=10.420,P=0.659)。模型实测发病趋势与预测发病趋势基本一致,预测值和实测值平均相对误差为23.88%,预测效果尚可。利用ARIMA(2,1,1)(1,1,1)12模型对2022年上城区其他感染性腹泻病发病情况进行预测,预测值均在95%置信限范围内,存在夏季和冬季两个发病高峰,和浙江省及全国监测数据相一致。结论ARIMA(2,1,1)(1,1,1)12模型能对上城区其他感染性腹泻病做出较好的预测,在早期防控中起到一定的作用。

Establishment and prediction of multiple seasonal ARIMA model for other infectious diarrhea in Shangcheng District

Objective To establish the multiple seasonal auto-regressive integrated moving average (ARIMA) model for other infectious diarrhea diseases in Shangcheng District, and to provide references for the early prevention and control. Methods SPSS 25.0 software was used to establish the multiple seasonal ARIMA model based on the data about the incidence of other infectious diarrhea diseases in Shangcheng District from 2010 to 2020. The fitting effect of the model was evaluated based on regression prediction on the number of monthly incidence in 2021, and the established model was used to predict the number of monthly incidence in 2022. Results A total of 40,534 cases of other infectious diarrhea diseases were reported in Shangcheng District from 2010 to 2020, with an average annual reported incidence of 3,685 cases. No death case was reported. The optimal model established was ARIMA (2,1,1) (1,1,1)₁₂, with stable R²=0.870, Bayesian information criterion (BIC)=9.524, mean absolute percentage error (MAPE)=27.351. No statistical significance was found in Box-Ljung test (Q=10.420, P=0.659). The measured incidence trend of the model was basically consistent with the predicted incidence trend, and the average relative error between the predicted value and the measured value was 23.88%, indicating a good prediction effect. ARIMA (2,1,1) (1,1,1)₁₂ model was employed to predict the incidence of other infectious diarrhea diseases in Shangcheng District in 2022. The predicted values were all within the 95% confidence limit, and there were two peaks of incidence in summer and winter, which wereconsistent with the monitoring data of Zhejiang Province and the whole country. Conclusion ARIMA (2,1,1) (1,1,1)₁₂ model can make a good prediction for other infectious diarrhea diseases in Shangcheng District, and plays a certain role in early prevention and control.