ARIMA模型在乌鲁木齐市儿童呼吸系统疾病发病趋势预测中的应用  

Application of ARIMA model in prediction of epidemic trend of respiratory diseases in children in Urumqi

作　　者：王艳杰[1] 杨蕾[1] 吴文华[2] 任泉[3] 王倩[1] 刘早玲[1] 

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医疗信息统计科,新疆乌鲁木齐830011 [3]乌鲁木齐市气象局业务科技处,新疆乌鲁木齐830011

出　　处：《职业与健康》2017年第21期2955-2959,共5页Occupation and Health

基　　金：新疆维吾尔自治区自然科学基金(2016D01C188)

摘　　要：目的应用时间序列分析方法建立差分自回归移动平均模型,也叫求和自回归移动平均模型(Autoregressive Integrated Moving Average Model,ARIMA),分析和预测2011年1月—2015年12月乌鲁木齐市儿童呼吸系统疾病的月门诊人次并作短期预测,为儿童呼吸系统疾病的预防和控制提供支持。方法以2011年1月—2015年12月乌鲁木齐市儿童呼吸系统疾病的月门诊人次为原始序列,用ARIMA(p,d,q)(P,D,Q)S拟合序列,根据自相关图(ACF)和偏自相关图(PACF)对模型定阶并估计参数,再对模型及其参数做显著性检验,应用最小信息量准则AIC及SBC进行评价,建立最优ARIMA乘积季节模型。结果 ARIMA(0,1,2)(1,1,0)_(12)模型模拟2011年1月—2015年6月儿童呼吸系统疾病月门诊人次,计算得平均绝对百分比误差MAPE=10.91,在可接受的范围内。运用该模型预测出2015年7月—12月的儿童呼吸系统疾病月门诊人次,MAPE=11.39,模型预测效果较好。该模型预测2016年全年的儿童呼吸系统疾病月门诊人次,最大值出现在12月,预测月门诊人次为8 963(5 431~12 494)。结论 ARIMA(0,1,2)(1,1,0)_(12)模型可用于拟合并且短期预测乌鲁木齐市儿童呼吸系统疾病月门诊人次,为儿童呼吸系统疾病的预防和治疗提供依据。[Objective] To establish the Autoregressive Integrated Moving Average Model （ARIMA） by time series analysis method,analyze and predict the monthly outpatient visits of children with respiratory diseases in Urumqi from January 2011 to December 2015 ,make short-term predictions,and provide a support for the prevention and control of respiratory diseases in chil- dren.[Methods] Based on the original series of monthly outpatient visits of children with respiratory diseases in Urumqi from January 2011 to December 2015, ARIMA （p, d, q ）（P, D, Q ）s was applied in fitting sequence, order determination and parameter estimation were performed according to ACF and PACF,then the significance test was conducted in the model and parameters, and the evaluation was performed by akaike information criterion （AIC） and SBC,to establish the optimal model of ARIMA. [ Results ] ARIMA （0, l, 2）（ 1,1,0）12 was used to simulate the monthly outpatient visits of children with respiratory diseases in Urumqi from January 2011 to December 2015 ,the average absolute percentage error （MAPE=10.91） was in an acceptable range. The model was used to predict the monthly outpatient visits of children with respiratory diseases from June to December in 2015, MAPE=I 1.39, which the prediction effect was better. Finally, the model was used to predict the monthly outpatient visits of children with respiratory diseases in 2016, the maximum occurred in December, and the monthly outpatient visits was 8 963 （5 431-12 494 ）.[ Conclusion ] ARIMA （0, 1,2） （1,1,0）12 can be used to fit and short-term predict monthly outpatient visits of children with respiratory diseases in Urumqi, provide basis for the prevention and treatment of respiratory diseases in children.

关 键 词：呼吸系统疾病 时间序列分析 ARIMA乘积季节模型 儿童 

分 类 号：R181.2[医药卫生—流行病学]