SARIMA模型在医院门诊量预测中的应用  

Applications of SARIMA Model on Predicting Outpatients Quantity

作　　者：李小升[1] 马春柳[1] 雷海科[1] 刘海霞[1] 

机构地区：[1]重庆市肿瘤研究所,重庆市400030

出　　处：《中国病案》2013年第3期37-40,共4页Chinese Medical Record

摘　　要：目的了解医院2000年-2011年门诊量的趋势,建立预测月门诊量的时间序列模型,为优化医疗资源配置提供科学的统计学依据。方法根据医院2002年1月至2012年12月年门诊量数据,应用SPSS18.0软件建立季节自回归滑动平均模型(Seasonal Auto Regressive Integrated Moving Average model,SARIMA模型),并验证2012年7至12月份的门诊量。结果预测值与实际值的上升下降趋势基本吻合,SARIMA(0,1,1)(0,1,1)12模型为最优模型,标准化贝叶斯信息标准(NormalizedBayesian Information Criteria,BIC)值与平均绝对误差百分比(Mean absolute percentage error,MPE)值最小,BIC值为13.82,MPE为7.70,Box-Ljung检验无统计学意义(Q18=17.93,P=0.3281>0.05)。结论 SARIMA模型能很好的拟合门诊量的变动趋势,在无外界因素影响的情况下,医院门诊量将会继续上涨。Objective To investigate outpatients quantity trend during 2002 and 2012, establish forecast month outpatients quan- tity time series model. In order to provide scientific statistics basis for the allocation of medical resources. Methwls Seasonal Autorcgressive Integrated Moving Average （SARIMA） models have been developed on the monthly data collected from January 2002 to Decomber 2012 and validated using the data from July 2012 to December 2012 by SPSS18.0 software. Results The predicted values were consistent with the upturns and downturns of the observed series. The SARIMA（0, 1,1）（0, 1, 1）12 model has been found as the most suitable model with least Normalized Bayesian Information Criteria （BIC） of 13.82 and Mean Absolute Percent Error （MAPE） of 7.70. The model was further validated by Ljung-Box test （ Q18 = 17.93, P = 0.3281 〉 0.05）. Conclusion SARIMA mode match perfectly outpatients changing trend and without external intervention factor, the outpatient quantity in our hospital will continually increase.

关 键 词：时间序列 SARIMA模型 门诊量 预测 

分 类 号：R197.3[医药卫生—卫生事业管理]