ARIMA季节乘积模型在预测医院月住院量的应用.pdfVIP

重庆医科大学学报2008年第33卷第4期 (Journal of Chongqing Medical University 2008．Vo1．33 No．4) 一 475一 文章编号：0253—3626(2008)04—0475—03 ARIMA季节乘积模型在预测医院月住院量的应用 刘金莲，钟晓妮 (重庆医科大学公共卫生学院卫生统计教研室，重庆 400016) 【摘 要】目的：探讨ARIMA季节乘积模型在时间序列资料分析中的应用，建立住院量的预测模型。方法：利用重庆市某医院 2002～2005年的月住院量，采用条件最／bS-乘法估计模型参数，按照残差不相关原则、简洁原则确定模型结构，依据AIC和SBC 准则确定模型阶数，建立ARIMA预测模型。结果：模型参数有统计学意义。方差估计值为0．003 175，AIC=一94．245 1，SBC= 一 91．252 1。对模型进行白噪声残差分析，拟合优度结果表明ARIMA的最优模型为：(1一B)(1一B )zF(1-0．543 02B)(1一B )at。结 论：用所建立模型对月住院量进行预测，结果表明ARIMA是一种短期预测精度较高的预测模型。 【关键词】时间序列；预测；自回归求和移动平均模型 【中国图书分类法分类号】R197．323 【文献标识码】A 【收稿日期】2007—09—05 Applications of multiple seasonal ARIMA model on predictive workload of inpatient department LIU矗n—lion，et al (Department of Stat~tics，College of Public Health，Chongqing Medical University) 【Abstract】Objective：To discuss the application of multiple seasonal AR1MA model to data in time series and fit predictive model on workload of inpatient department．Methods：Parameter of model is estimated using conditional least squares method according to the data of workload of inpatient department in a hospital in Chongqing from 2002 to 2005．The structure is determined according to criteria of residual un-correlation and concision．ARIMA predictive model was fitted and the order of model was confirmed through Akaike’S Inform ation Criterion and Schwarz Bayesian Criterion ． Results：Statistics assisted estimation of the mode1．Th e estimation of variance is 0．003 175，AIC=-94．245 1，SBC=一91．252 1．Using analysis of white—noise residual of model，fit of rich table show that the best estimated ARIMA model is (1-B)(1-B )zt=(1-0．543 02B)(1-W2)a~． Conclusion：The model of ARIMA can be used to forecast for workload of i