深圳市ARIMA在肾综合征出血热发病预测中应用  

Application of autoregressive integrated moving average model in incidence rate prediction of hemorrhagic fever with renal syndrome in Shenzhen city

作　　者：梅树江[1] 周志峰[2] 马汉武[1] 周海涛[2] 樊丹怡[3] 王敬忠[1] 李媛[1] 李怀昕[1] 柯跃斌[1] 

机构地区：[1]深圳市疾病预防控制中心传染病防制科,广东518055 [2]深圳市福田区疾病预防控制中心 [3]深圳市妇幼保健院

出　　处：《中国公共卫生》2015年第7期936-938,共3页Chinese Journal of Public Health

基　　金：深圳市科技研发资金知识创新计划基础研究项目(JCYJ20130329103949657)

摘　　要：目的建立时间序列分析的自回归求和移动平均(ARIMA)模型,预测深圳市肾综合征出血热(HFRS)发病趋势。方法深圳市2005—2013年HFRS逐月发病率建立预测深圳市HFRS的最优ARIMA模型,利用2014年逐月HFRS发病率回代来检验模型预测效果,根据预测值与实际值的相对误差判断模型的预测精度,再以2005—2014年HFRS逐月发病率构建模型预测2015年的HFRS发病率。结果模型ARIMA(1,0,1)(1,0,1)12较好地拟合既往时间段的发病序列,各项参数(AR=0.993,MA=0.926,SAR=0.967,SMA=0.857)均有统计学意义(P<0.01),BIC值=-3.300,Ljung-Box模型统计量Q=20.794,P=0.107,模型残差为白噪声,2014年逐月HFRS发病率的预测值符合实际值的变动趋势,全年发病率预测值与实际值的相对误差率为20.74%。预测2015年深圳市HFRS发病率为4.28/100万。结论 ARIMA模型能很好地模拟深圳市HFRS发病率在时间序列上的变化趋势,并对未来的发病率进行预测。Objective To establish an autoregressive integrated moving average（ ARIMA） model for epidemic trend prediction of hemorrhagic fever with renal syndrome（ HFRS） in Shenzhen city. Methods ARIMA model was established based on the monthly incidence rate of HFRS in Shenzhen city from 2005 through 2013. The constructed optimal model was applied to predict the incidence rate of HFRS in Shenzhen city in 2014 and to evaluate the validity of the model through comparing the difference of predicted incidence rate and actual one. The incidence rate of HFRS in 2015 was predicted by the ARIMA model based on the incidence rate from 2005 through 2014. Results The model of ARIMA（ 1,0,1）（ 1,0,1）12had a good fitness to the incidence rate and all parameters（ AR = 0. 993,MA = 0. 926,SAR = 0. 967,SMA = 0. 857） of the model were statistically significant（ P〈0. 01） and Schwarz Bayesian criterion（ BIC） =- 3. 300. The residual error was white noise and the Ljung-Box test statistics for the model was 20. 794,with a P value of 0. 107. The predicted incidence rate in 2014 was consistent with the actual one,with the relative error of20. 74%. The predicted incidence rate of HFRS in 2015 based on the incidence rate from 2005 through 2014 is 4. 28 per billion. Conclusion ARIMA model can be used to simulate changes in HFRS incidence rate and to forecast the incidence rate in future in Shenzhen city.

关 键 词：肾综合征出血热 时间序列 ARIMA模型 预测 

分 类 号：R373.3[医药卫生—病原生物学]