摘要
目的观察求和自回归移动平均(ARIMA)乘积季节模型在我院门诊量预测中的应用,为我院卫生资源配置决策提供合理依据。方法收集我院2009年~2018年逐月门诊量数据,使用R语言构建ARIMA乘积季节模型,预测2019年月门诊量,并用预测值与实际值的平均绝对百分误差(MAPE)评价预测效果。结果ARIMA(1,1,2)×(2,1,0)12模型是我院门诊量相对最佳预测模型,模型残差经Ljung-Box检验证实为白噪声序列(Q=16.126,P=0.1856);模型预测值与实际值的MAPE为6.56%,均在预测值的95%可信区间内,模型预测精度较高。结论ARIMA(1,1,2)×(2,1,0)12乘积季节模型能较好地应预测我院门诊量,医院应根据门诊量变化规律合理配置人力资源及医疗物资,增强现代医院的门诊综合服务能力及效率。
Objective To observe the application of the sum-autoregressive moving average(ARIMA)product seasonal model in the prediction of outpatient volume in our hospital,and to provide a reasonable basis for the decision on the allocation of health resources in our hospital.Methods Collecting monthly outpatient volume data in our hospital from 2009 to 2018,use the R language to construct an ARIMA product seasonal model to predict the monthly outpatient volume in 2019,and use the average absolute percentage error(MAPE)of the predicted value and the actual value to evaluate the prediction effect.Results The ARIMA(1,1,2)×(2,1,0)12 model was the relatively best predictive model for outpatient volume in our hospital.The model residuals were confirmed to be white noise series by Ljung-Box test(Q=16.126,P=0.1856);The MAPE between the predicted value of the model and the actual value was 6.56%,which were both within the 95%confidence interval of the predicted value,and the prediction accuracy of the model was high.Conclusion The ARIMA(1,1,2)×(2,1,0)12 product seasonal model can better predict the outpatient volume in our hospital.Hospitals should rationally allocate human resources and medical supplies according to the law of changes in outpatient volume,and enhance the comprehensive outpatient service capabilities and efficiency of modern hospitals.
作者
唐路
宋萍
谢冰珏
佘颖
TANG Lu;SONG Ping;XIE Bing-jue;SHE Ying(Department of Medical Record Statistics,Children's Hospital of Chongqing Medical University,Chongqing 400014,China)
出处
《医学信息》
2021年第11期19-22,共4页
Journal of Medical Information
关键词
门诊量
ARIMA乘积季节模型
资源配置
人力资源
医疗物资
Outpatient volume
ARIMA product seasonal model
Resource allocation
Human resources
Medical supplies