指数平滑法与ARIMA模型在四级手术人次预测中的应用  

Application of exponential smoothing model and ARIMA model in the prediction of the fourth-level surgeries

作　　者：桂成 王国林[1] 陶源[1] 曹冬梅[1] GUI Cheng;WANG Guolin;TAO Yuan;CAO Dongmei(The First People’s Hospital of Changzhou/The Third Affiliated Hospital of Soochow University,Changzhou 213003,China)

机构地区：[1]常州市第一人民医院//苏州大学第三附属医院,江苏常州213003

出　　处：《现代医院》2021年第12期1860-1863,共4页Modern Hospitals

摘　　要：目的对某三甲医院的四级手术人次建立时间序列模型,利用指数平滑法和ARIMA模型对其进行评价及预测。方法从某三甲医院的统计报表中摘取2014年1月—2018年12月的四级手术人次,分别采用指数平滑法和ARIMA对其进行建模,利用2019年1—12月四级手术人次数据验证模型,比较两种模型的拟合及预测效果。结果运用指数平滑法构建的最优模型是Winters加法模型,模型的平稳R^(2)为0.723,MAPE为5.092,标准BIC为7.941。ARIMA模型筛选出的最优模型为ARIMA(0,1,1)(1,1,0)_(12),模型的平稳R^(2)为0.553,MAPE为6.564,标准BIC为8.781。结论 Winters加法模型比ARIMA模型拟合及预测效果更好,可以为医院管理决策提供更为科学的理论依据。Objective A time series was established for the fourth-level surgeries in a tertiary hospital, and the exponential smoothing method and ARIMA method were used to establish corresponding models to predict the time series. Methods The exponential smoothing method and ARIMA method were used to model the number of four-level surgeries extracting from January 2014 to December 2018 from the statistical reports of a tertiary hospital. Using the fourth-level surgeries data from January to December 2019 to verify the models by comparing the fitting and prediction effects of the two models. Results The optimal model constructed by the exponential smoothing method is the Winters additive model. The stationary R^(2) square of the model is 0.723, MAPE is 5.092 and BIC is 7.941. The optimal model selected by ARIMA model is ARIMA(0,1,1)(1,1,0)_(12). The stationary R^(2) square of the model is 0.553, MAPE is 6.564, and BIC is 8.781. Conclusion Winters additive model is better than ARIMA model in fitting and predicting effect, and it provides scientific theoretical basis for hospital management decision.

关 键 词：指数平滑法 ARIMA模型 四级手术人次 预测 

分 类 号：R197.323[医药卫生—卫生事业管理]