ARIMA模型在江西省布鲁氏菌病发病数预测中的应用

目的 构建江西省布病流行趋势预测模型,为布病防控提供科学依据。方法 利用2014-2017年江西省人间布病月发病数建立不同差分次数下的最优模型,同时,对2018年1月的布病发病数进行预测,比较模型的拟合、预测效果。结果 不同差分条件的最优模型分别为AR(1)、ARIMA(1,1,3)、ARIMA(3,2,0)、ARIMA(3,3,0)。其中,模型AR(1)对2018年1月的发病数的预测误差最小。结论 ARIMA模型用于江西省布病发病数的短期预测基本可行。

ARIMA model in prediction of brucellosis incidence in Jiangxi Province,China

In order to build a prediction model of brucellosis epidemic trend in Jiangxi Province,this study used the monthly incidence of human brucellosis in Jiangxi Province from 2014 to 2017 to establish an optimal model under the condition of different differential times periods.At the same time,each model was used to predict the incidence of brucellosis in January 2018,and compared the fitting and prediction effects for each model.The results show that the optimal models of different difference conditions are AR(1),ARIMA(1,1,3),ARIMA(3,2,0)and ARIMA(3,3,0)respectively,Among which,the prediction error of model AR(1)on the number of cases in January 2018 was the smallest.It is indicated that the short-term prediction of brucellosis incidence in Jiangxi Province by ARIMA model is basically feasible.