Connection between multilayer perceptrons and regression using independent component analysis

The data model of independent component analysis (ICA) gives a multivariate probability density that describes many kinds of sensory data better than classical models like Gaussian densities or Gaussian mixtures. When only a subset of the random variables is observed, ICA can be used for regression,...

Full description

Saved in:
Bibliographic Details
Published in:Neurocomputing (Amsterdam) 2003, Vol.50, p.211-222
Main Authors: Hyvärinen, Aapo, Bingham, Ella
Format: Article
Language:English
Subjects:
Independent component analysis
Multilayer perception
Nonlinear regression
Projection pursuit
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:The data model of independent component analysis (ICA) gives a multivariate probability density that describes many kinds of sensory data better than classical models like Gaussian densities or Gaussian mixtures. When only a subset of the random variables is observed, ICA can be used for regression, i.e. to predict the missing observations. In this paper, we show that the resulting regression is closely related to regression by a multi-layer perceptron (MLP). In fact, if linear dependencies are first removed from the data, regression by ICA is, as a first-order approximation, equivalent to regression by MLP. This theoretical result gives a new interpretation of the elements of the MLP: The outputs of the hidden layer neurons are related to estimates of the values of the independent components, and the sigmoid nonlinearities are obtained from the probability densities of the independent components.
ISSN:0925-2312
1872-8286
DOI:10.1016/S0925-2312(01)00705-6