Multiclass linear dimension reduction by weighted pairwise Fisher criteria

Top Cited Papers

1 July 2001

journal article
Published by Institute of Electrical and Electronics Engineers (IEEE)

Vol. 23 (7) , 762-766
https://doi.org/10.1109/34.935849

Abstract

We derive a class of computationally inexpensive linear dimension reduction criteria by introducing a weighted variant of the well-known K-class Fisher criterion associated with linear discriminant analysis (LDA). It can be seen that LDA weights contributions of individual class pairs according to the Euclidean distance of the respective class means. We generalize upon LDA by introducing a different weighting function.

Keywords

This publication has 8 references indexed in Scilit:

Multi-class linear feature extraction by nonlinear PCA
Published by Institute of Electrical and Electronics Engineers (IEEE) ,2002
Heteroscedastic discriminant analysis and reduced rank HMMs for improved speech recognition
Speech Communication, 1998
The nonlinear PCA learning rule in independent component analysis
Neurocomputing, 1997
Toward Bayes-optimal linear dimension reduction
Published by Institute of Electrical and Electronics Engineers (IEEE) ,1994
CANONICAL VARIATE ANALYSIS—A GENERAL MODEL FORMULATION
Australian Journal of Statistics, 1984
A formulation and comparison of two linear feature selection techniques applicable to statistical classification
Pattern Recognition, 1984
Feature combinations and the divergence criterion
Computers & Mathematics with Applications, 1977
THE STATISTICAL UTILIZATION OF MULTIPLE MEASUREMENTS
Annals of Eugenics, 1938