基于指数平滑模型的克拉玛依市流感样病例预测分析  

Predictive analysis of influenza-like illness based on exponential smoothing method in Karamay city

作　　者：蔺洁[1] 李霞[1] 刘佳[1] LIN Jie;LI Xia;LIU Jia(Kalamay Municipal Center for Disease Control and Prevention,Karamay,Xinjiang 834000,China)

机构地区：[1]新疆克拉玛依市疾病预防控制中心,新疆克拉玛依834000

出　　处：《疾病预防控制通报》2021年第6期8-11,共4页Bulletin of Disease Control & Prevention(China)

基　　金：克拉玛依市创新人才工程(2019RC001A-03)。

摘　　要：目的应用指数平滑法预测新疆克拉玛依市流感样病例就诊比(ILI%)变化趋势,为当地流感防控工作提供技术支持。方法利用2018年1 w—2020年52 w的ILI%数据拟合建立指数平滑模型,选择最优模型预测2021年1-26 w ILI%,并以实际值和预测值的相对误差验证模型预测效果。结果 2018—2020年克拉玛依市ILI%呈季节性周期变化;简单季节模型为最优模型,平均相对误差为43.63%,2021年1-9 w平均相对误差为21.83%,第八周相对误差最小为3.62%,10-26 w平均相对误差为55.18%。结论指数平滑模型对ILI%的短期预测效果较好。Objective To fit and predict the incidence trend of influenza-like illness ratio(ILI%) with exponential smoothing method in Karamay city of Xinjiang, and provide technical support for local ILI prevention and control. Methods The fitting of weekly ILI% data from week 1, 2018 to week 52, 2020 were taked for establishing exponential smoothing models. The optimal model was selected to predict the ILI% of from week 1 to 26, 2021, and the effect of prediction model was verified by the relative errors between the actual values and the predictive values. Results The incidence rates of influenza-like illness showed seasonal variation, and the simple seasonal model was optimal with the average relative error of 43.63% from 2018 to 2020. The mean of relative errors was 21.83% from week 1 to 9, 55.18% from week 10 to 26, and the relative error was minimum(3.62%) in week 8, 2021. Conclusions The effect of exponential smoothing model is good on short-term prediction of ILI epidemic.

关 键 词：指数平滑法 流感 预测 

分 类 号：R511.7[医药卫生—内科学]