电子学报2012，Vol.40Issue(2)：358-364,7.DOI:10.3969/j.issn.0372-2112.2012.02.024
基于类别保留投影的基因表达数据特征提取新方法
New Method of Feature Extraction for Gene Expression Data Based on Class Preserving Projection
摘要
Abstract
A new method of discriminant feature extraction,called Class Preserving Projection (CPP) ,is proposed from the point view of class relation of pairwise samples. Compared to LDA, CPP has the following two advantages. One is that the optimal subspace dimension is not restricted to the number of categories of samples, and the other is that computational complexity is lower. Experiments are performed on gene expression data for sample classification, and the results confirm the effectiveness of the method. Kernel CPP (KCPP) is presented by generalizing CPP to nonlinear space to solve the problem of nonlinear feature extraction, and the experiments on gene expression data verify the feasibility of the method.关键词
特征提取/fisher线性鉴别分析/小样本/基因表达数据Key words
feature extraction/Fisher's linear discriminant analysis(LDA)/small sample size/gene expression data分类
信息技术与安全科学引用本文复制引用
王文俊..基于类别保留投影的基因表达数据特征提取新方法[J].电子学报,2012,40(2):358-364,7.基金项目
中央高校基本科研业务费专项资金(No.K5051203013) （No.K5051203013）
