摘要
目的建立求和自回归移动平均(autoregressive integrated moving average,ARIMA)模型,预测岳阳市流感样病例(influenza-like illness,ILI)的发病趋势,为当地流感防控提供依据。方法收集岳阳市2010年第1周—2018年第10周的ILI发病资料,构建ARIMA模型,对岳阳市流感发病情况进行预测和评价。结果2010—2018年岳阳市ILI周发病数逐年缓慢上升,未见明显固定的季节性趋势。基于2010—2017年的ILI周发病数,建立ARIMA(1,1,4)模型,预测2018年前10周的ILI发病数,实际值均落入预测值95%置信区间,预测效果较好。结论ARIMA模型可以较好地对岳阳市流感样病例的发病趋势进行短期预测,而长期预测须不断加入新的数据调整模型参数。
Objective To predict the incidence trend of influenza-like illness(ILI)in Yueyang using autoregressive moving average(ARIMA)model,which could provide a scientific basis for the prevention and control of influenza.Methods The data of ILI from the 1 st week to the 10 th week in Yueyang were collected and analyzed using ARIMA model to forecast and evaluate the incidence of ILI.Results The weekly cases of ILI in Yueyang increased slowly year by year.There was not obvious and regular seasonal trend.Based on weekly ILI cases in 2010-2017,an ARIMA(1,1,4)model was established to predict the incidence of ILI in the first 10 weeks of 2018.95%CI of predicted values contained all actual values,which meant the forecast effect was good.Conclusion ARIMA model could predict the incidence trend of ILI in Yueyang well in a short time.However,long-term prediction must continuously add new data to adjust the model parameters.
作者
余艳妮
廖青
聂绍发
魏祥
娄常兴
YU Yanni;LIAO Qing;NIE Shaofa(Medical College,Yueyang Vovational Technical College,Yueyang,Hunan,414000,Chin)
出处
《中国社会医学杂志》
2020年第6期650-653,共4页
Chinese Journal of Social Medicine
基金
湖南省自然科学基金青年基金项目(2019JJ50632)。
关键词
流感样病例
监测
ARIMA模型
预测
Influenza-like illness
Surveillance
ARIMA model
Forecast
