An analysis of the GLVQ algorithm

Author

Gonzalez, A.I. ; Grana, Manuel ; d´Anjou, A.

Author_Institution

Dept. CCIA, Pais Vasco Univ., San Sebastian, Spain

Volume

6

Issue

4

fYear

1995

fDate

7/1/1995 12:00:00 AM

Firstpage

1012

Lastpage

1016

Abstract

Generalized learning vector quantization (GLVQ) has been proposed in as a generalization of the simple competitive learning (SCL) algorithm. The main argument of GLVQ proposal is its superior insensitivity to the initial values of the weights (code vectors). In this paper we show that the distinctive characteristics of the definition of GLVQ disappear outside a small domain of applications. GLVQ becomes identical to SCL when either the number of code vectors grows or the size of the input space is large. Besides that, the behavior of GLVQ is inconsistent for problems defined on very small scale input spaces. The adaptation rules fluctuate between performing descent and ascent searches on the gradient of the distortion function

Keywords

neural nets; sensitivity analysis; unsupervised learning; vector quantisation; ascent search; code vectors; descent search; generalized learning vector quantization; input space; neural nets; sensitivity; simple competitive learning; Algorithm design and analysis; Entropy; Gaussian distribution; Impedance matching; Neural networks; Organizing; Proposals; Stochastic processes; Sufficient conditions; Vector quantization;

fLanguage

English

Journal_Title

Neural Networks, IEEE Transactions on

Publisher

ieee

ISSN

1045-9227

Type

jour

DOI

10.1109/72.392266

Filename

392266