ARIMA与SARIMA模型在医院门诊人次与出院人次预测中的应用

Application of ARIMA and SARIMA model in predicting hospital outpatients and discharges

目的应用ARIMA与SARIMA模型对某院门诊人次与出院人次进行分析和预测。方法收集某医院2011年1月至2016年6月各月份的门诊人次和出院人次数据,运用ARIMA与SARIMA模型对数据进行分析,并验证2016年7~10月数据;预测2016年11月至2017年6月的门诊人次和出院人次。使用统计软件SPSS17.0进行模型拟合与预测。结果 ARIMA(6,1,0)是该院门诊人次的最优拟合预测模型,SARIMA(0,1,1)(0,1,0)_(12)(不含常数项)是该院出院人次的最优拟合预测模型,两模型均能较好地拟合2016年7月至10月的门诊人次与出院人次,用该模型预测2016年11月至2017年6月的数据,符合门诊人次与出院人次的变动趋势,实际值都在预测的95%上下限范围之内。结论ARIMA与SARIMA模型对短期门诊人次与出院人次预测具有较强的实用价值,可以为医院的卫生资源配置和科学管理提供参考依据。

Objective The ARIMA and SARIMA model were used to analyze and predict the number of outpatients and discharges. Methods We collected outpatient and discharge numbers in a hospital each month from January 2011 to June 2016, used ARIMA and SARIMA model to analyze data, and validated data from July 2016 to October 2016 to predict outpatients and discharges from November 2016 to June 2017. SPSS 17.0statistical software was used for model fitting and forecast. Results The ARIMA （6, O） was the optimal fitting model in predicting hospital outpatients, and the SARIMA, 1,1 （0） , （0, O） （exclu- ding constant） was the optimal fitting model in predicting hospital discharges. Both models could be well fitted in predicting out- patients and discharges from July 2016 to October 2016. The prediction with the two models from November 2016 to June 2017 conformed to the change trend of outpatients and discharges, and the actual values fell in the 95% upper and lower limits of pre- diction. Conclusion The ARIMA and SARIMA models have strong practical value in prediction of short-term outpatients and discharges, and can provide the reference to hospital health resource allocation and scientific management.