Stability of Oja's PCA Subspace Rule

1 July 1994

journal article
Published by MIT Press in Neural Computation

Vol. 6 (4) , 739-747
https://doi.org/10.1162/neco.1994.6.4.739

Abstract

This paper deals with stability of Oja's symmetric algorithm for estimating the principal component subspace of the input data. Exact conditions are derived for the gain parameter on which the discrete algorithm remains bounded. The result is extended for a nonlinear version of Oja's algorithm.

Keywords

INPUT DATA
PCA SUBSPACE
OJA'S ALGORITHM
SYMMETRIC ALGORITHM
PAPER DEALS
NONLINEAR VERSION
OJA'S PCA
SUBSPACE RULE

This publication has 4 references indexed in Scilit:

Least mean square error reconstruction principle for self-organizing neural-nets
Neural Networks, 1993
Convergence analysis of local feature extraction algorithms
Neural Networks, 1992
Principal components, minor components, and linear neural networks
Neural Networks, 1992
NEURAL NETWORKS, PRINCIPAL COMPONENTS, AND SUBSPACES
International Journal of Neural Systems, 1989