Sparse LDA

This is the MATLAB implementation of an algorithm for classification with FLDA designed for the high-dimensional/small sample size setting, exploiting among others sparsity. It is described in detail in our paper

,,Improving Implementation of Linear Discriminant Analysis for the High Dimension/Small Sample Size Problem"

with P. Schlesinger, published in Computational Statistics and Data Analysis, vol. 52, no.1, pp. 423--437, 2007. The software has been integrated in the Biosig software package, under GPL-licence. It can be found under the following link.

Generating matrices and initial vectors with arbitrary Ritz values and GMRES residual norms

These are the MATLAB subroutines belonging to our paper

,,Any Ritz value behavior is possible for Arnoldi and for GMRES with any convergence curve"

with G. Meurant. They can be found under the following link.

A look-ahead C-step for the Minimum Covariance Determinant estimator

These are the MATLAB subroutines belonging to our submitted paper

,,A computationally inexpensive modification of the C-step for the Minimum Covariance Determinant estimator"

with J. Kalina. They can be found under the following link.

Generating matrices and initial vectors with arbitrary harmonic Ritz values and residual norms for GMRES

These are the MATLAB subroutines belonging to our paper

,,Any admissible harmonic Ritz value set is possible for GMRES"

with K. Du and G. Meurant. They can be found under the following link.