逐步判别分析法在基因表达数据分类中的应用  

Application of Stepwise Discriminant Analysis In Classifying the Gene Expression Data

作　　者：邹杨[1] 陈忠[1] 谢俊宇 

机构地区：[1]长江大学信息与数学学院,湖北荆州434023 [2]洪湖贺龙高级中学,湖北洪湖433200

出　　处：《长江大学学报（自然科学版）》2011年第1期18-19,278-279,共2页Journal of Yangtze University(Natural Science Edition)

摘　　要：利用逐步判别分析法对基因表达值数据进行研究,筛选出了能够反映正常人和癌症患者在基因表达值上差异的特征基因。然后,以这些特征基因的表达值为训练样本建立Bayes判别函数,对未知类型的基因表达值序列进行了判别分类。结果表明,利用筛选出的变量进行判别时,其错判概率要远远小于利用所用变量进行判别时的情况。使用SAS分析系统完成以上步骤。In this paper,the gene expression data are studied by using the stepwise discriminant analysis method.By using this method,some special genes are screened out .They can reflect the difference between the healthy and the cancered persons. Then, based on the training samples, the Bayes linear discrimination functions are established.The samples was made up of the expression data of the special genes. By using these functions,the gene expression data are classified whose type were unknown.The results show that when the discrimination function is established by the selected variables, its error rate is lower than that built by all the gene expression data.All the steps are processed by SAS system .

关 键 词：逐步判别分析 基因表达水平 分类 

分 类 号：O221.2[理学—运筹学与控制论]