GM(1,1)残差修正模型在乙肝发病预测中的应用  

Application of GM(1,1) Modified Residual Error Model in the Prediction of Hepatitis B Prevalence

作　　者：李滨[1] 张卫民 

机构地区：[1]辽阳市疾病预防控制中心,111000 [2]辽阳市太子河区疾控站

出　　处：《中国卫生统计》2014年第2期240-242,共3页Chinese Journal of Health Statistics

摘　　要：目的了解和掌握辽阳市乙肝发病规律和流行趋势,为乙肝防治提供科学依据。方法应用GM(1,1)残差修正模型对《中国疾病预防控制信息系统》中辽阳市2004-2012年乙肝发病率数据拟合并进行动态预测。结果传统GM(1,1)预测模型为^x(1)k+1=(x(0)1-μ/α)e-αk+μ/α=1983.4029e0.0479k-1844.0793,精度检验结果显示模型精度勉强合格;GM(1,1)残差修正模型:^x(1)2(k+1)=1983.4029e0.0479k-736.925e-0.054k-196.6992e0.113k-933.4641,均能通过残差检验和关联度检验判定标准,后验差检验其模型精度判别为优;预测2013年和2014年辽阳市乙肝发病率分别为134.0638/10万和138.5362/10万。结论 GM(1,1)残差修正模型拟合辽阳市乙肝发病率理想,预测精度高,具有一定的实用性和精确性。Objective To understand and master the regularity and epidemic trends of virus hepatitis B in Liaoyang city for providing the scientific basis for prevention and treatment of virus hepatitis B. Methods To apply the GM （ I, 1 ） modified residual error model to fit and forcast dynamically the data of hepatitis B incidence of China Information System for Disease Control and Prevention in Liaoyang city from 2004 to 2012. Results It showed the traditional GM （ 1,1 ） prediction model was x（1）k＋1 = （x（0）1 -μ/α）e-αk μ/α = 1983.4029e0.079k - 1844. 0793 and the result of model precision-test was only basic. The GM （ 1,1 ） modified residual error model was x（1）2 （ k ＋ 1 ） = 1983. 4029e0.0479k - 736. 925e-0.054k - 196. 6992e0.113k - 933. 4641 and Can reach criteria for residual test and correlation test, in addition the model was advanced by the precision-test. The predictive values of the incidence of Hepatitis B were 134. 0638/105 and 138. 5362/105 respectively during 2013 and 2014. Conclusion The GM（ 1,1 ） modified residual error model had a good fitness and hizh orecision .but also a certain practicalitv and accuracy.

关 键 词：乙型病毒性肝炎 GM(1 1) 残差修正 模型 预测 

分 类 号：R512.62[医药卫生—内科学]