组合模型对肺结核发病趋势的预测研究  

Combination Model for Predicting the Incidence of Pulmonary Tuberculosis

作　　者：陈银苹[1] 吴爱萍[2] 余亮科[3] 许雅丽[1] 蒋宁[1] 杨阳[3] 张锦[3] 张静宇[3] 曹燕花[1] 

机构地区：[1]河北联合大学公共卫生学院,河北省唐山市063000 [2]河北联合大学公共卫生学院病理系,河北省唐山市063000 [3]河北联合大学附属医院,河北省唐山市063000

出　　处：《中国全科医学》2014年第21期2452-2456,共5页Chinese General Practice

基　　金：唐山市科学技术发展课题(131302156);河北联合大学青年科学研究基金资助项目(z201225)

摘　　要：目的建立肺结核发病率(1/10万)自回归积分移动平均(ARIMA)-灰色模型(GM)组合模型,并将其应用于肺结核发病率的预测,为及早发现疾病发展趋势和及时采取控制对策提供科学依据。方法收集迁安市2004年1月—2012年12月肺结核月发病率资料,应用SPSS 13.0软件对肺结核逐月发病率进行ARIMA建模拟合;然后用GM(1,1)模型对其带阈值的残差序列进行修正并构造出组合模型,利用此模型对迁安市2013年肺结核逐月发病率进行预测。结果 ARIMA(0,1,1)(0,1,1)12模型较好地拟合了肺结核的发病情况,模型的所有参数都通过统计学检验;用一阈值为4的GM(1,1)模型对其残差序列进行修正,预测模型通过了精度检验(C=0.573,P=0.805),模型拟合精度为基本合格,ARIMA-GM组合模型的平均绝对误差(MAE)、平均绝对百分比误差(MAPE)都比单个模型小,利用组合模型对2013年肺结核发病率预测。结论 ARIMA-GM组合模型能较好地拟合迁安市肺结核发病情况,且该方法比ARIMA季节乘积模型预测具有更高的精度。预测结果能够对肺结核的早期预测预警模型的建立提供借鉴,从而有针对性地采取相应的控制措施。ObjectiveTousetheautoregressiveintegratedmovingaverage（ARIMA）andthegreymodel（GM） combination model to predict pulmonary tuberculosis incidence in Qian′an, so as to provide scientific basis for decision making in the early warning system of infectious diseases .Methods The incidence of pulmonary tuberculosis was collected monthly between January 2004 and December 2012 and ARIMA model was analyzed with SPSS 13.0 statistical software.Then, the GM （1, 1） model with a threshold value of its residual sequence correction was used .Finally, the results of the incidence of pulmonary tuber-culosis was used to establish the combination model to predict the monthly incidence in 2013.Results The ARIMA （0, 1, 1） （0, 1, 1）12 model fitted well the incidence of pulmonary tuberculosis , and all the parameters underwent statistical tests .The GM （1, 1） model with a threshold of 4 was used to correct its residuals , and the forecasting model underwent accuracy test （ C=0.573, P=0.805） and the fitting accuracy was qualified and can be used for extrapolation .The MAE and MAPE of the ARIMA-GM combination model were both smaller than a single model .Conclusion The ARIMA-GM combined model can better fit the incidence of tuberculosis in Qian′an city, and can be used to forecast the incidence of tuberculosis in Qian′an so as to provide evidence foor the establishing of early warning system and to adopt corresponding measures .Additionally , the combined model has a greater accuracy than the seasonal ARIMA model .

关 键 词：结核 肺 预测 自回归积分移动平均模型 灰色模型 

分 类 号：R521[医药卫生—内科学]