ARIMA模型在三门峡市手足口病发病预测中的应用

目的 建立以三门峡市手足口病发病数据为基础的整合滑动平均自回归模型(autoregressive integrated moving average model, ARIMA),并利用该模型对三门峡市手足口病的发病数进行预测。 方法 以三门峡市2008年1月—2017年12月的手足口病月发病数据为基础差分平稳化后经过文献查阅和验证建立最优ARIMA,并对2018年1月—12月的手足口病发病数进行预测,通过与实际值的比较评价预测效果。 结果 三门峡市手足口病发病预测模型为ARIMA(1,1,1)×(0,1,1)₁₂ ,模型各项参数均有统计学意义(P<0.001),拟合优度检验BIC=0.287,残差序列为白噪声序列(P=0.10),拟合效果较好。预测了三门峡市2018年1—12月的手足口发病数,并与实际值进行比较,1—2月预测值与实际值符合度较高。 结论 拟合的三门峡市手足口病发病序列模型ARIMA(1,1,1)×(0,1,1)₁₂效果较好,可用于对三门峡市手足口病发病趋势进行短期预测。

Application of autoregressive integrated moving average model to predicting the incidence of hand,foot and mouth disease in Sanmenxia city

Objective To establish an autoregressive integrated moving average model(ARIMA)based on the incidence data of hand,foot and mouth disease(HFMD)in Sanmenxia city,and use the model to predict the incidence of HFMD in Sanmenxia city.Methods Based on the monthly incidence data of HFMD in Sanmenxia city from January 2008 to December 2017,the optimal ARIMA model was established after literature review and verification.The number of HFMD cases from January to December 2018 was predicted,and the predictive effect was evaluated by comparing with the actual value.Results The forecasting model of HFMD incidence in Sanmenxia city was ARIMA(1,1,1)×(0,1,1)12,and all parameters of the model had statistical significance(P<0.001).Through the test of goodness of fit,the value of Bayesian information criterion(BIC)was 0.287,the residual sequence was a white noise sequence(P=0.10),and the fitting effect was better.The number of HFMD cases in Sanmenxia city from January to December 2018 was predicted,and then a comparison of the predicted and actual values was conducted.A higher degree of conformity was found between the predicted and actual values from January to February.Conclusions The fitting effect of the ARIMA model of HFMD incidence in Sanmenxia city is good,and it can be used to predict the short-time incidence trend of HFMD in Sanmenxia city.