Compressive Regularized Discriminant Analysis of High-Dimensional Data with Applications to Microarray Studies

TL;DR

This paper introduces CRDA, a novel method for high-dimensional data analysis that improves gene selection and classification accuracy in microarray studies by combining regularization and joint-sparsity techniques.

Contribution

CRDA is a new discriminant analysis method that effectively performs feature elimination and gene selection in high-dimensional microarray data.

Findings

CRDA outperforms competitors in reducing misclassification errors.

CRDA achieves accurate feature selection in microarray datasets.

Simulation and real data demonstrate CRDA's effectiveness.

Abstract

We propose a modification of linear discriminant analysis, referred to as compressive regularized discriminant analysis (CRDA), for analysis of high-dimensional datasets. CRDA is specially designed for feature elimination purpose and can be used as gene selection method in microarray studies. CRDA lends ideas from $\ell_{q,1}$ norm minimization algorithms in the multiple measurement vectors (MMV) model and utilizes joint-sparsity promoting hard thresholding for feature elimination. A regularization of the sample covariance matrix is also needed as we consider the challenging scenario where the number of features (variables) is comparable or exceeding the sample size of the training dataset. A simulation study and four examples of real-life microarray datasets evaluate the performances of CRDA based classifiers. Overall, the proposed method gives fewer misclassification errors than its competitors, while at the same time achieving accurate feature selection.