贝叶斯判别分析在布氏杆菌常见种别鉴定中的应用  

The Application of Bayes Discriminant Analysis in the Identification of Brucella's Species

作　　者：肖培[1] 崔步云[1] 

机构地区：[1]传染病预防控制国家重点实验室感染性疾病诊治协同创新中心中国疾病预防控制中心传染病预防控制所102206

出　　处：《中国卫生统计》2013年第6期802-804,共3页Chinese Journal of Health Statistics

基　　金：125重大专项课题(2013ZX10004221,2012ZX10004215);973课题资助(2010CB530201)

摘　　要：目的建立布氏杆菌常见种别的Bayes判别模型,为布氏杆菌种别的鉴定提供依据。方法以多位点可变数目串联重复序列分析方法所选择的基因组上16个具有多态性串联重复序列位点为判别指标,应用逐步判别分析方法筛选关键序列位点并构建布氏杆菌常见种别的Bayes判别模型。结果待选的16个多态性串联重复序列位点全部进入判别函数,建立的模型自身验证的准确率为98.21%,交互验证的准确率为97.75%。结论建立的Bayes判别模型判断布氏杆菌的常见种别正确率很高,是布氏杆菌常见种别判断的有效方法。Objective To establish a Bayes discriminant mod- el and provide scientific basis for the identification of the popular species of brucella. Methods Using stepw ise method to select the VNTR from mul- tiple locus variable number tandem repeats analysis as the discriminant vari- ables and Bayes method to build discriminant model. Results All of the sixteen VNTR w ere selected in the model. 98. 21% of original grouped ca- ses w ere correctly classified and 97. 75% of cross-validated grouped cases w ere correctly classified. Conclusion The established Bayes discrimi- nant model has a high level of accuracy and is an useful method for the judgment of the popular species of Brucella.

关 键 词：布氏杆菌 Bayes判别模型 多位点可变数目串联重复序列分析 

分 类 号：R516.7[医药卫生—内科学]