Abstract:Linear discriminant analysis (LDA) has been a useful tool in pattern recognition and data analysis research and practice. While linearity of class boundaries cannot always be expected, nonlinear projections through pre-trained deep neural networks have served to map complex data onto feature spaces in which linear discrimination has served well. The solution to binary LDA is obtained by eigenvalue analysis of within-class and between-class scatter matrices. It is well known that the multiclass LDA is solved by an extension to the binary LDA, a generalised eigenvalue problem, from which the largest subspace that can be extracted is of dimension one lower than the number of classes in the given problem. In this paper, we show that, apart from the first of the discriminant directions, the generalised eigenanalysis solution to multiclass LDA does neither yield orthogonal discriminant directions nor maximise discrimination of projected data along them. Surprisingly, to the best of our knowledge, this has not been noted in decades of literature on LDA. To overcome this drawback, we present a derivation with a strict theoretical support for sequentially obtaining discriminant directions that are orthogonal to previously computed ones and maximise in each step the Fisher criterion. We show distributions of projections along these axes and demonstrate that discrimination of data projected onto these discriminant directions has optimal separation, which is much higher than those from the generalised eigenvectors of the multiclass LDA. Using a wide range of benchmark tasks, we present a comprehensive empirical demonstration that on a number of pattern recognition and classification problems, the optimal discriminant subspaces obtained by the proposed method, referred to as GO-LDA (Generalised Optimal LDA), can offer superior accuracy.

A New Weighted Way of Linear Discriminant Function

Computational and Theoretical Analysis of Supervised Dimensionality Reduction

Constrained Linear Discrimination Analysis and Face Recognition

Varying Coefficient Linear Discriminant Analysis for Dynamic Data

Beyond Trace Ratio: Weighted Harmonic Mean of Trace Ratios for Multiclass Discriminant Analysis

A Note on Two-Dimensional Linear Discriminant Analysis.

A New Reduced-Rank Linear Discriminant Analysis Method and Its Applications

A Generalized Form of Fisher Linear Discriminant Function

Linear discriminant analysis based on genetic algorithm

Discriminative analysis algorithm for linear local tangent spaces

Optimal Linear Discriminant Analysis for High-Dimensional Functional Data

A procedure of linear discrimination analysis with detected sparsity structure for high-dimensional multi-class classification

Linear discriminant analysis with spectral regularization

Linear Discriminant Analysis Using Rotational Invariant L1 Norm

Communications Inspired Linear Discriminant Analysis

Network linear discriminant analysis

Linear Discriminant Regularized Regression

GO-LDA: Generalised Optimal Linear Discriminant Analysis

A Discriminant Model for the Pattern Recognition of Linearly Independent Samples

An efficient method for computing orthogonal discriminant vectors

A Direct Estimation Approach to Sparse Linear Discriminant Analysis