Abstract:Linear discriminant analysis (LDA) represents a simple yet powerful technique for partitioning a p-dimensional feature vector into one of K classes based on a linear projection learned from N labeled observations. However, it is well-established that in the high-dimensional setting (p > N) the underlying projection estimator degenerates. Moreover, any linear discriminate function involving a large number of features may be difficult to interpret. To ameliorate these issues, two general categories of sparse LDA modifications have been proposed, both to reduce the number of active features and to stabilize the resulting projections. The first, based on optimal scoring, is more straightforward to implement and analyze but has been heavily criticized for its ambiguous connection with the original LDA formulation. In contrast, a second strategy applies sparse penalty functions directly to the original LDA objective but requires additional heuristic trade-off parameters, has unknown global and local minima properties, and requires a greedy sequential optimization procedure. In all cases the choice of sparse regularizer can be important, but no rigorous guidelines have been provided regarding which penalty might be preferable. Against this backdrop, we winnow down the broad space of candidate sparse LDA algorithms and promote a specific selection based on optimal scoring coupled with a particular, complementary sparse regularizer. This overall process ultimately progresses our understanding of sparse LDA in general, while leading to targeted modifications of existing algorithms that produce superior results in practice on three high-dimensional gene data sets.

Towards Robust and Sparse Linear Discriminant Analysis for Image Classification

Efficient and Robust Sparse Linear Discriminant Analysis for Data Classification

Generalized robust linear discriminant analysis for jointly sparse learning

Low-rank discriminative least squares regression for image classification

Robust sparse manifold discriminant analysis

Understanding and Evaluating Sparse Linear Discriminant Analysis

Self-weighted Robust LDA for Multiclass Classification with Edge Classes

Robust Linear Discriminant Analysis Using Ratio Minimization of L1,2-Norms

Triple Sparse Denoising Discriminantive Least Squares Regression for Image Classification

Laplacian Sparse Dictionary Learning for Image Classification Based on Sparse Representation

Constrained Linear Discrimination Analysis and Face Recognition

Denoising Low-Rank Discrimination based Least Squares Regression for image classification

Regularized Deep Linear Discriminant Analysis

Linear Discriminant Regularized Regression

Regularisation Constrained Denoising Discriminant Least Squares Regression for Image Classification

Optimal Dimensionality Discriminant Analysis and Its Application to Image Recognition

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

Double Constrained Discriminative Least Squares Regression for Image Classification

3E-LDA

L2,1-Norm Regularized Robust and Sparse Linear Discriminant Analysis via an Alternating Direction Method of Multipliers

Toward Robust Discriminative Projections Learning Against Adversarial Patch Attacks