季节时间序列分析在新疆地区涂阳结核疫情预测预警中的应用  

Application of seasonal time series analysis to forecasting and early warning of epidemics of smear-positive tuberculosis in Xinjiang region

作　　者：张伟文[1] 贺湘焱[2] 古丽娜扎尔·艾克拜尔 曹明芹[1] ZHANG Wei-wen;HE Xiang-yan;AIKEBAIER·Gulinazhaer;CAO Ming-qin(School of Public Health, Xinjiang Medical University, Xinjiang, Urumqi 830011, China;Xinjiang Uygur Municipal People’s Hospital, Xinjiang, Urumqi 830001, 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]新疆维吾尔自治区人民医院,新疆乌鲁木齐830001

出　　处：《实用预防医学》2019年第1期26-29,共4页Practical Preventive Medicine

基　　金：新疆维吾尔自治区自然科学基金/联合基金资助(2016D01C167)

摘　　要：目的 探讨季节时间序列模型(seasonal autoregressive integrated moving average,SARIMA)在新疆涂阳肺结核疫情预测预警中的应用。 方法 收集新疆2005年1月-2015年6月涂阳肺结核月发病数建立时间序列分析模型,对2015年7-12月的月发病数进行预测并与实际发病数进行比较评价。 结果 新疆涂阳肺结核月发病数具有明显季节性,在3月和11月存在发病高峰,SARIMA(0,1,1)(0,1,1)12模型参数的回归系数假设检验结果P<0.001,与实际月发病数拟合效果良好,平均绝对百分比误差MAPE=7.985%。 结论 SARIMA模型能较好的拟合新疆涂阳肺结核疫情在时间上的变异趋势,用其进行预测效果良好,能够为肺结核疫情的预警防治提供参考。Objective To explore the application of seasonal autoregressive integrated moving average (SARIMA) model in the prediction and early warning of epidemics of smear-positive tuberculosis in Xinjiang. Methods The monthly incidence of smear-positive tuberculosis in Xinjiang from January 2005 to June 2015 was collected to establish a time series analysis model. The monthly incidence from July to December in 2015 was predicted and compared with the actual incidence in the same period. Results The monthly incidence of smear-positive tuberculosis in Xinjiang showed obvious seasonal variations, and the incidence peaked in March and November. Hypothesis test result showed that the regression coefficient of the SARIMA (0,1,1)(0,1,1)12 model parameters was P<0.001, and the model had a good fitting effect for actual monthly incidence. The mean absolute percentage error (MAPE) was 7.985%. Conclusions The SARIMA model demonstrates goodness-of-fit in forecasting the changing trend of epidemics of smear-positive tuberculosis in Xinjiang, which provides references for early warning and prevention of tuberculosis epidemics.

关 键 词：时间序列 涂阳结核 疫情预警 

分 类 号：R183.3[医药卫生—流行病学]