应用SARIMA模型预测医院出院人次  

Prediction on the Number of Discharged Patients with the Application of SARIMA Mode

作　　者：李秋燕[1] 王津雨[1] 

机构地区：[1]中国医学科学院北京协和医学院血液病医院(血液学研究所),天津市300020

出　　处：《中国病案》2016年第7期32-35,共4页Chinese Medical Record

摘　　要：目的对医院出院人次进行预测分析,旨在为医院科学管理提供参考依据。方法对某院2006年1月-2014年12月的出院人数进行SARIMA模型的建模拟合,利用2015年1月-12月的出院人数对模型进行检验,然后利用所建立的模型预测2016年的出院人数。结果 SARIMA(2,1,1)(0,1,1)12(不含常数项)是该院出院人数的最优拟合预测模型,模型的R2=0.956,贝叶斯信息准则(BIC)为9.05;模型的平均相对误差绝对值为3.60%;2016年的出院人数预测值为22111。结论出院人数具有季节变动及长期增长趋势,SARIMA模型对于短期的医院出院人数预测具有较强的实用价值,可以为科学管理提供参考依据。Objective To conduct prediction analysis on the number of discharged patients, and provide reference basis for scientific management of the hospital. Methods To establish SARIMA mode with the number of discharged patients in a hospital from January 2006 to December 2014, conduct inspection on the number of discharged patients from January to December in 2015, then predict the discharged patient number in 2016 with the mode established. Results SARIMA（2,1,1）（0,1,1）12（constant term not included） was the best prediction mode of the discharged patients number of our hospital, R2=0.956, BIC was 9.05, the absolute value of average relative error of the discharged patients number was 3.60%, the prediction number of 2016 was 22111. Conclusions The number of discharged patients had a tendency of seasonal variation and longtime increase, the SARIMA mode had relatively high practical value on the prediction of discharged patient number in short time, and provide evidence for the scientific management of the hospital.

关 键 词：SARIMA模型 时间序列 出院人数 预测 

分 类 号：R197.323[医药卫生—卫生事业管理]