一种基于监督降维和形状分析的基因选择方法  

Gene selection method based on supervised dimension reduction and procrustes analysis

作　　者：耿耀君[1] 张军英[1] 

机构地区：[1]西安电子科技大学计算机学院,陕西西安710071

出　　处：《西安电子科技大学学报》2011年第3期121-127,共7页Journal of Xidian University

基　　金：国家自然科学基金资助项目(61070137);国家自然科学基金重点资助项目(60933009);陕西省科技攻关资助项目(2009K01-56)

摘　　要：因为由主分量分析与形状分析相结合的基因选择方法没有有效利用样本的类别信息,所以提出了一种新的基因选择方法,将间隔最大化判别分析和形状分析相结合,在选择基因过程中不仅考虑了基因与基因之间的相互作用,也考虑了基因与类之间的相互关系,提高了所选基因集的分类性能.对4组微阵列基因表达数据的实验表明,该方法的性能优于主分量分析与形状分析相结合的方法,与当前两个流行的多变量F ilter方法相比,也具有一定的优势.The gene selection method which combines principal component analysis with shape analysis does not effectively use the class information on samples. Aiming at this shortcoming, a new gene selection method combining margin maximizing discriminant analysis with shape analysis is presented in this paper. In the gene selection process, the new method considers not only the interaction between genes but also the relationship between genes and class label, which improves the classification performance of selected genes. Experimental results on four mieroarray gene expression data show that the performance of the presented method is superior to that of the method which combines principal component analysis with shape analysis. Compared with two state- of-the-art muhivariable filter methods, the presented method also has a certain advantage.

关 键 词：基因选择 微阵列数据 Procrustes分析 间隔最大化判别分析 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]