ARIMA模型在预测重庆市医院日住院量中的应用

目的建立预测与监测的求和自回归移动平均模型(ARIMA)的时间序列模型,研究日住院量的变化规律。方法通过对2009年2～4月重庆市逐日住院患者量分析用Box-Ljung统计量评价ARIMA模型的拟合度,用平均预测相对误差作为预测效果的评价指标。结果重庆市住院患者量以周为时间周期,每周中以周一、二住院量达到高峰,周六、日为低谷。ARIMA(0,1,1)(1,1,1)7是重庆市2009年2～4月住院量预测最优拟合预测模型,一周和两周外推预测的平均相对误差分别为6.27%和9.14%。结论对住院患者量的历史数据进行时间序列分析是用于住院患者量监测的一个重要的内容。本研究所建立的ARIMA模型适用于重庆市住院患者量预测,预测精度较高。

Application of ARIMA model on predicting everyday workload of inpatient department of hospitals in Chongqing

Objective To develop the autoregressive integrated moving-average(ARIMA)time series model for forecasting and monitoring the everyday workload of inpatient department in Chongqing for studying its change law.Methods Statistics of Box-Ljung was used to evaluate the degree of fitness of ARIMA model,and the average relative errors of prediction were used as indexes to evaluate the predict effect.Results The changes of everyday workload of inpatient department in Chongqing presented a weekly periodicity,and showed that everyday workload of inpatient department from Monday to Tuesday exceeded its weekly average.The model ARIMA(0,1,1)(0,1,1)7 was the best fitted model to predict the inpatient workload of Chongqing from February to April in 2009.The average relative errors of predicts in one week and two weeks were 6.27% and 9.14%.Conclusion The time series method applied to the historical reporting data of everyday workload of inpatient department is an important tool for everyday workload of inpatient department surveillance.The ARIMA model is suitable to forecast everyday workload of inpatient department in Chongqing with high forecasting accuracy.