Abstract:Linear discriminant analysis (LDA) is a classical method for dimensionality reduction, where discriminant vectors are sought to project data to a lower dimensional space for optimal separability of classes. Several recent papers have outlined strategies, based on exploiting sparsity of the discriminant vectors, for performing LDA in the high-dimensional setting where the number of features exceeds the number of observations in the data. However, many of these proposed methods lack scalable methods for solution of the underlying optimization problems. We consider an optimization scheme for solving the sparse optimal scoring formulation of LDA based on block coordinate descent. Each iteration of this algorithm requires an update of a scoring vector, which admits an analytic formula, and an update of the corresponding discriminant vector, which requires solution of a convex subproblem; we will propose several variants of this algorithm where the proximal gradient method or the alternating direction method of multipliers is used to solve this subproblem. We show that the per-iteration cost of these methods scales linearly in the dimension of the data provided restricted regularization terms are employed, and cubically in the dimension of the data in the worst case. Furthermore, we establish that when this block coordinate descent framework generates convergent subsequences of iterates, then these subsequences converge to the stationary points of the sparse optimal scoring problem. We demonstrate the effectiveness of our new methods with empirical results for classification of Gaussian data and data sets drawn from benchmarking repositories, including time-series and multispectral X-ray data, and provide Matlab and R implementations of our optimization schemes.

Sparse Exponential Discriminant Analysis

Sparse Exponential Discriminant Analysis and Its Application to Fault Diagnosis

Generalized Discriminant Analysis: A Matrix Exponential Approach

Understanding and Evaluating Sparse Linear Discriminant Analysis

A Direct Estimation Approach to Sparse Linear Discriminant Analysis

Optimal Feature Selection for Sparse Linear Discriminant Analysis and Its Applications in Gene Expression Data

Inexact Implementation Using Krylov Subspace Methods for Large Scale Exponential Discriminant Analysis with Applications to High Dimensionality Reduction Problems.

A Simple and Effective Generalization of Exponential Matrix Discriminant Analysis and Its Application to Face Recognition.

SRDA: an Efficient Algorithm for Large-Scale Discriminant Analysis

SRDA: an Efficient Algorithm for Large-Scale

Sparse tensor discriminant analysis

Sparse L1-norm-based linear discriminant analysis

Proximal methods for sparse optimal scoring and discriminant analysis

Speed up kernel discriminant analysis

$\mathcal{DBSDA}$ : Lowering the Bound of Misclassification Rate for Sparse Linear Discriminant Analysis via Model Debiasing

Efficient and Robust Sparse Linear Discriminant Analysis for Data Classification

Robust sparse manifold discriminant analysis

3E-LDA

Sparse linear discriminant analysis by thresholding for high dimensional data

Sparse Multiple Maximum Scatter Difference for Dimensionality Reduction

Sparse ordinal discriminant analysis