应用自回归移动平均模型预测医院出院人数  

Application of AMIAR Model in the Prediction of Discharged Patients' Number

作　　者：黄利娟[1] 梁学柱[1] 查君敬 

机构地区：[1]安庆市立医院,安庆市246003

出　　处：《中国病案》2014年第9期35-36,34,共3页Chinese Medical Record

摘　　要：目的 探讨自回归移动平均模型（ARIMA）在某院出院人数预测中的应用,从而为医院的科学管理决策服务.方法 对某院2000年-2013年的月出院人数数据进行收集,将2000年-2012年的数据用于建立ARIMA模型,2013年数据用于验证所建立的模型,所建模型预测2014年出院人数,统计软件采用SPSS17.0.结果 建立的ARIMA（0,1,1）（0,1,1）12模型具有较高的拟合度,预测的2013年出院人数各月相对误差为0.27％～15.68％,全年出院人数平均相对误差为4.86％,预计2014年出院人数为68,880人次.结论 ARIMA模型适用于出院人数的预测,对于出院人数变化规律的分析有较好的适应性,但在预测远期数据时则应综合多方面因素.Objective To explore the application of AMIAR model in the prediction of discharged patients＇ number of a hospital, and provide scientific management decision service for the hospital. Methods To collect the data of discharged patients＇ number of a hospital monthly from 2000 to 2013 and established the AMIAR model with the data which collected from 2000 to 2012. The data of 2013 was used to verify the model, then made a prediction of the discharged patients number of 2014 with the model, all statistic analysis was conducted by SPSS 17.0 software. Results The ARIMA （0,1,1） （0,1,1）12 model had a high fitting degree. The absolute relative error for the months of 2013 ranged from 0.27% to 15.68%, and the average relative error for whole year was 4.86%. The prediction of discharged patients＇ number in 2014 was 68880 cases. Conclusions The ARIMA model was applicable to the prediction of discharged patients＇ number and had well adaptation to the analysis of the variation regularity, however, we should take comprehensive factors into consideration in the prediction of long-term data.

关 键 词：ARIMA模型 时间序列 出院人数 预测 

分 类 号：R197.32[医药卫生—卫生事业管理]