Lasso based gene selection for linear classifiers
Selecting a subset of genes with strong discriminative power is a very important step in classification problems based on gene expression data. Lasso is known to have automatic variable selection ability in linear regression analysis. This paper uses Lasso to select most informative genes to represent the class label as a linear function of the gene expression data. The selected genes are further used to fit linear classifiers for tumor classification. The proposed approach (gene selection and linear classification) was applied to 5 publicly available cancer datasets. Compared to other methods in literature, the proposed method achieves similar or higher classification accuracy with fewer genes.
Cross validation, Lasso, Leave-one-out, Linear classifier, Variable selection
Zheng, Songfeng, and Weixiang Liu. "Lasso based gene selection for linear classifiers." In 2009 IEEE International Conference on Bioinformatics and Biomedicine Workshop, pp. 203-208. IEEE, 2009.
Proceedings - 2009 IEEE International Conference on Bioinformatics and Biomedicine Workshops, BIBMW 2009