ARIMA模型在合肥市疟疾发病预测中的应用  

Application of ARIMA model in the forecasting of malaria incidence in Hefei city

作　　者：樊雯婧[1] 陆群[2] 邹立巍[1] 仰凤桃[2] 王波[2] 武松[3] 翟金霞[1] 

机构地区：[1]安徽医科大学公共卫生学院劳动卫生与环境卫生学系,合肥230032 [2]合肥市疾病预防控制中心消地科,合肥230061 [3]安徽中医学院中西医结合临床学院,合肥230038

出　　处：《安徽医科大学学报》2013年第3期252-256,共5页Acta Universitatis Medicinalis Anhui

基　　金：合肥市2010年度第一批科技计划项目(编号:[2011]25号)

摘　　要：目的探讨自回归求和移动平均(ARIMA)模型预测合肥市疟疾发病率的可行性,并预测疟疾月发病趋势。方法应用SPSS 13.0软件对合肥市疟疾逐月发病率建立ARI-MA模型,利用2011年疟疾月发病率对模型参数进行校正从而预测2012年疟疾月发病率。结果模型ARIMA(1,1,1)(1,1,0)12是合肥市疟疾拟合的最佳模型,预测值与实际值基本吻合,且实际值均在预测值95%可信区间范围内,其外推预测的平均相对误差为2.57%。结论用时间序列模型对疟疾发病情况的拟合结果满意,预测效果良好,可用于预测未来疟疾的变动趋势,为疟疾预防控制措施的制定提供重要依据。Objective Using the ARIMA （Autoregressive Integrated Moving Average） model to predict the monthly incidence trend of malaria in Hefei city. Methods To construct the multiple seasonal ARIMA model based on the monthly malaria incidence of Hefei city from 1999 to 2011 with SPSS 13.0 software. To modify the parameters of the ARIMA model according to the monthly malaria incidence in 2011 and then forecasted the monthly malaria incidence in 2012 in Hefei city. Results The ARIMA （1, 1, 1 ） （1,1,0） 12 model was the best model to forecast the malaria incidence in Hefei city. The prediction vaules of malaria incidence fit the practical situation. All true values fell in the 95 % confidence interval of expected values for expected case numbers of malaria by the model. The mean relative error of the model was 2.57 %. Conclusion It is practical to apply the approach of ARIMA product season model to predicting malaria incidence in Hefei city by testing and the model seemed to be an appropriate model to predict the future incidence trend. This ARIMA model could provide evidence for the development of malaria prevention and control measures.

关 键 词：ARIMA模型 时间序列分析 疟疾 预测 发病率 

分 类 号：R531.3[医药卫生—内科学]