乙型肝炎发病趋势的时间序列分析和预测模型研究  

APPLICATION OF TIME SERIES ANALYSIS IN THE PREDICTION OF INCIDENCE TREND OF HEPATITIS B

作　　者：钱燕华[1] 林玉娣[1] 姚健[1] 彭志行[2] 顾静[1] 陈峰[2] 

机构地区：[1]江苏省无锡市疾病预防控制中心,无锡214023 [2]南京医科大学流行病与卫生统计学系

出　　处：《现代预防医学》2009年第14期2606-2609,共4页Modern Preventive Medicine

基　　金：江苏省预防医学科研基金(Y2006003);2006年度无锡市科技局科技发展计划重大建设项目(CLE00617);2006年度无锡市社会发展科技指导性计划(CLZ00632)

摘　　要：[目的]利用乘积季节模型预测无锡市乙型肝炎(乙肝)的发病趋势,为制订相应的防治措施提供科学依据。[方法]利用最小二乘原理,应用自回归求和移动平均模型与随机季节模型相结合的乘积季节模型,对无锡市1998～2006年乙肝月发病数进行时间序列分析并建立预测模型,用2007年相应数据验证预测效果,并对2008年以后无锡乙肝发病趋势进行预测。[结果]利用1998～2006年资料构建ARIMA(1,1,1)(1,1,1)12模型,所建立模型的预测效果良好,实际值均在预测值的95%可信区间内,预测2008～2012年无锡乙肝发病人数呈下降趋势。[结论]采用ARIMA乘积季节模型预测无锡市乙肝发病情况,拟合及近期预测效果均较好。预测结果符合我市乙肝发病现状及目前采取的乙肝防治措施。[ Objective ] To utilize the product seasonal Model to predict the trend of hepatitis B, so as to provide scientific basis for formulating the related measures of prevention and cure. [ Methods] The product seasonal Model was combined with ARIMA model and stochastic seasonal model by using the least square principle. The incidence of hepatitis B from 1998 to 2006 in Wuxi was analyzed by time series analysis to establish the prediction model. The data of 2007 was used to verified the effectiveness of this model, and forecasted and analyzed the trends of incidence of hepatitis B in Wuxi after 2008 years. [ Resuits] The ARIMA（ 1,1,1 ） （ 1,1,1 ）^12 model was established by the data from 1998 to 2006, and the effectiveness of prediction of this model showed to be good with the actual values in the 95% confidence interval of predicted values. The study showed that the trends of incidence of hepatitis B would decrease from 2008 to 2012. [ Conclusion] The ARIMA product seasonal model shows effective to predict the incidence of hepatitis B in Wuxi, and the results is in according to the current status of hepatitis B, moreover, it could provide information for us to take measures for hepatitis B prevention and control.

关 键 词：乙型肝炎 时间序列分析 乘积季节模型 预测 

分 类 号：R512.62[医药卫生—内科学]