On the equivalence of Kernel Fisher discriminant analysis and Kernel Quadratic Programming Feature Selection

Authors:
I. Rodriguez-Lujan;C. Santa Cruz;R. Huerta
Affiliations:
Departamento de Ingenierıa Informática and Instituto de Ingenierıa del Conocimiento, Universidad Autónoma de Madrid, 28049 Madrid, Spain;Departamento de Ingenierıa Informática and Instituto de Ingenierıa del Conocimiento, Universidad Autónoma de Madrid, 28049 Madrid, Spain;BioCircuits Institute, University of California, San Diego, La Jolla, CA 92093-0404, USA
Venue:
Pattern Recognition Letters
Year:
2011

Citing 4
Cited 1

Nonlinear component analysis as a kernel eigenvalue problem

Neural Computation
Computer-Aided Multivariate Analysis

Computer-Aided Multivariate Analysis
Kernel Methods for Pattern Analysis

Kernel Methods for Pattern Analysis
Quadratic Programming Feature Selection

The Journal of Machine Learning Research

Analysis of pattern recognition and dimensionality reduction techniques for odor biometrics

Knowledge-Based Systems

Quantified Score

Hi-index	0.10

Visualization

Abstract

We reformulate the Quadratic Programming Feature Selection (QPFS) method in a Kernel space to obtain a vector which maximizes the quadratic objective function of QPFS. We demonstrate that the vector obtained by Kernel Quadratic Programming Feature Selection is equivalent to the Kernel Fisher vector and, therefore, a new interpretation of the Kernel Fisher discriminant analysis is given which provides some computational advantages for highly unbalanced datasets.