ARIMA季节乘积模型在儿童肺炎门急诊人次预测中的应用  

Application of ARIMA seasonal product model to predicting the number of outpatient and emergency department visits for childhood pneumonia

作　　者：杨蕾[1] 吴文华[2] 任泉[3] 王艳杰[1] 刘早玲[1] YANG Lei;WU Wen-hua;REN Quan;WANG Yan-jie;LIU Zao-ling(School of Public Health,Xinjiang Medical University,Xinjiang,Urumqi 830054,China;Center for Medical Information,the First Affiliated Hospital of Xinjiang Medical University,Urumqi ,Xinjiaag 830011,China;Ururnqi Weather Bureau,Urumqi,Xinjiang 830002,China)

机构地区：[1]新疆医科大学公共卫生学院,新疆乌鲁木齐830011 [2]新疆医科大学第一附属医院信息中心,新疆乌鲁木齐830054 [3]乌鲁木齐市气象局,新疆乌鲁木齐830002

出　　处：《实用预防医学》2019年第1期33-36,共4页Practical Preventive Medicine

基　　金：新疆维吾尔自治区自然基金(2016D01C188);新疆医科大学科研创新基金(XYDCX201512)

摘　　要：目的 探索自回归差分移动平均(autoregressive integrated moving average,ARIMA)季节乘积模型在预测儿童肺炎门急诊人次的应用,为合理利用医疗资源提供科学依据。 方法 收集乌鲁木齐市两家三级甲等医院2011-2016年儿童肺炎逐月门急诊人次数据,使用R 3.4.1软件进行模型的识别、参数估计与检验,建立ARIMA季节乘积模型对2011年1月-2016年6月儿童肺炎逐月门急诊人次进行拟合,并利用2016年7-12月数据计算预测值与实际值的平均预测相对误差来评价预测效果。 结果 ARIMA(0,1,2)(1,1,0)12模型是拟合儿童肺炎门急诊人次的最佳预测模型,平均相对误差为9.82%。 结论 ARIMA 季节乘积模型有较好的拟合和短期预测效果,能为医院合理利用医疗资源提供参考依据。Objective To explore the application of autoregressive integrated moving average (ARIMA) seasonal product model to forecasting the number of outpatient and emergency department visits for childhood pneumonia, and to provide a scientific basis for the rational utilization of medical resources. Methods We collected the data regarding the number of monthly outpatient and emergency department visits for childhood pneumonia from two tertiary comprehensive hospitals in Urumqi City from 2011 to 2016. R 3.4.1 software was used for model identification, parameter estimation and verification. ARIMA seasonal product model was established to fit the number of monthly outpatient and emergency department visits for childhood pneumonia from January 2011 to June 2016, and the prediction effect was evaluated through checking the average relative error between the predicted and actual values by using the monthly data from July to December in 2016. Results The ARIMA (0,1,2) (1,1,0)12 model was the best prediction model, with an average relative error of 9.82%. Conclusions ARIMA seasonal product model has better fitting and short-term prediction effect, which can provide references for hospitals to rationally utilize medical resources.

关 键 词：儿童肺炎 门急诊人次 ARIMA季节乘积模型 预测 

分 类 号：R181.2[医药卫生—流行病学]