ARIMA模型与GRNN模型对性病发病率的预测研究  

Prediction on the incidence of blood and sexually transmitted diseases with models of ARIMA and GRNN

作　　者：韩琴[1] 苏虹[1] 王忱诚[1] 单晓伟[1] 常微微[1] 徐志伟[1] 王静[1] 韩红梅[2] 

机构地区：[1]安徽医科大学公共卫生学院流行病与卫生统计学系,安徽合肥230032 [2]安徽省合肥市疾病预防控制中心

出　　处：《现代预防医学》2012年第6期1337-1340,共4页Modern Preventive Medicine

基　　金：2009年度博士科研经费--猩红热发病率预测模型效果的比较研究资助项目(XJ200907)

摘　　要：目的比较ARIMA模型与GRNN模型在性病发病率预测上的有效性。方法以新站区2005～2009年传染病疫情资料为基础,利用月发病率建立ARIMA(0,0,1)模型;将各月的发病率作为GRNN模型的输入,对应下个月发病率作为网络的输出,对样本进行训练,比较两模型的应用效果。结果合肥市新站区2005～2009年性病平均发病率达135.32/10万,并以年平均发展速度1.03缓慢上升。ARIMA(0,0,1)模型与GRNN模型的平均误差率分别是33.8%和27.0%;R2值分别是0.714和0.749。结论新站区性病呈平缓的波浪式上升,GRNN模型较适合该类疾病的发病率预测,建议公共卫生人员依据疫情预测及时做好防控工作。OBJECTIVE To compare the prediction efficiency of sexually transmitted diseases with models of ARIMA and GRNN. METHODS ARIMA （0, 0, 1 ） model with the incidence of each month was established, the incidence of each month as GRNN model was input, and the corresponding next month incidence as the network was output. The the application of two models was compared. RESULTS The incidence of sexually transmitted diseases was 135.32,/100 000, and rose slowly at aver- age annual rate of 1.03 during 2005 to 2009 in xinzhan district of Hefei. The mean error rates （MSE） of ARIMA model and GRNN model were 33.8% and 27.0%, respectively; the R2 values of two models were 0.714 and 0.749, respectively. CON- CLUSION The incidence of the sexual transmission disease showed a gentle wave-like uplift, GRNN model was suitable for forecasting the incidence of such diseases, and the result recommends the public health officer should control the epidemic situ- ation according to forecast results in time.

关 键 词：性病 流行病学方法 ARIMA模型 GRNN模型 

分 类 号：R195.1[医药卫生—卫生统计学]