Discriminative Dimensionality Reduction Based on Generalized LVQ

Authors:
Atsushi Sato
Affiliations:
-
Venue:
ICANN '01 Proceedings of the International Conference on Artificial Neural Networks
Year:
2001

Citing 5
Cited 2

Introduction to statistical pattern recognition (2nd ed.)

Introduction to statistical pattern recognition (2nd ed.)
Blind separation of sources, Part 1: an adaptive algorithm based on neuromimetic architecture

Signal Processing
A fast fixed-point algorithm for independent component analysis

Neural Computation
Prototype Learning Algorithms for Nearest Neighbor Classifier with Application to Handwritten Character Recognition

ICDAR '99 Proceedings of the Fifth International Conference on Document Analysis and Recognition
Pattern recognition using discriminative feature extraction

IEEE Transactions on Signal Processing

Maximal margin approach to kernel generalised learning vector quantisation for brain-computer interface

ICONIP'12 Proceedings of the 19th international conference on Neural Information Processing - Volume Part III
Comparing multi-objective and threshold-moving ROC curve generation for a prototype-based classifier

Proceedings of the 15th annual conference on Genetic and evolutionary computation

Quantified Score

Hi-index	0.01

Visualization

Abstract

In this paper, a method for the dimensionality reduction, based on generalized learning vector quantization (GLVQ), is applied to handwritten digit recognition. GLVQ is a general framework for classifier design based on the minimum classification error criterion, and it is easy to apply it to dimensionality reduction in feature extraction. Experimental results reveal that the training of both a feature transformation matrix and reference vectors by GLVQ is superior to that by principal component analysis in terms of dimensionality reduction.