Local Discriminant Analysis

Authors:
Marco Loog;Dick de Ridder
Affiliations:
Image Group, IT University of Copenhagen Copenhagen, Denmark;Delft University of Technology, Delft, The Netherlands
Venue:
ICPR '06 Proceedings of the 18th International Conference on Pattern Recognition - Volume 03
Year:
2006

Citing 0
Cited 4

On an alternative formulation of the Fisher criterion that overcomes the small sample problem

Pattern Recognition
Locally linear embedding based on correntropy measure for visualization and classification

Neurocomputing
A unified dimensionality reduction framework for semi-paired and semi-supervised multi-view data

Pattern Recognition
From the idea of "sparse representation" to a representation-based transformation method for feature extraction

Neurocomputing

Quantified Score

Hi-index	0.00

Visualization

Abstract

The main objective of the work presented here is to introduce a supervised, nonlinear dimensionality reduction technique which, performs well-known linear discriminant analysis in a local way and which is able to provide a powerful mapping with less computational effort than other nonlinear reduction methods. Additionally, because of the close connection of the new approach to Fisher's LDA, it is more clear that it acts discriminatively, which is not immediately apparent from previous formulations. The method makes use of the optimal scoring framework advocated by Hastie et al. and it is coined local discriminant analysis (eDA).