Non-linear ICA by Using Isometric Dimensionality Reduction

doi:10.1.007/978-3-540-30110-3_90

Non-linear ICA by Using Isometric Dimensionality Reduction

by John A Lee, Christian Jutten, Michel Verleysen

View in Mendeley Desktop In your library Save reference to library

Lecture Notes in Computer Science (2004)

Volume: 3195, Issue: December, Publisher: Springer, Pages: 710-717

DOI: 10.1.007/978-3-540-30110-3_90

Available from Lecture Notes in Computer Science

or

Abstract

In usual ICA methods, sources are typically estimated by maximizing a measure of their statistical independence. This paper ex- plains how to perform non-linear ICA by preprocessing themixtures with recent non-linear dimensionality reduction techniques. These techniques are intended to produce a low-dimensional representation of the data (the mixtures), which is isometric to their initial high-dimensional distribu- tion. A detailed study of the mixture model that makes the separation possible precedes a practical example.