Non-linear feature extraction by the coordination of mixture models

Jakob Verbeek; Nikos Vlassis; Ben Krose

Non-linear feature extraction by the coordination of mixture models

Jakob Verbeek ¹ Nikos Vlassis ¹ Ben Krose ¹

Abstract : We present a method for non-linear data projection that offers non-linear versions of Principal Component Analysis and Canonical Correlation Analysis. The data is accessed through a probabilistic mixture model only, therefore any mixture model for any type of data can be plugged in. Gaussian mixtures are one example, but mixtures of Bernoulli's to model discrete data might be used as well. The algorithm minimizes an objective function that exhibits one global optimum that can be found by finding the eigenvectors of some matrix. Experimental results on toy data and real data are provided.

Keywords : Canonical Correlation Analysis Principal Component Analysis Self-organizing Maps

Type de document :

Communication dans un congrès

9th Annual Conference of the Advanced School for Computing and Imaging (ASCI '03), Jun 2003, Heijen, Netherlands. pp.287--293, 2003

Domaine :

Informatique [cs] / Apprentissage [cs.LG]

Liste complète des métadonnées

Littérature citée [15 références] Voir Masquer Télécharger

https://hal.inria.fr/inria-00321490
Contributeur : Jakob Verbeek <>
Soumis le : mardi 8 mars 2011 - 15:04:10
Dernière modification le : lundi 25 septembre 2017 - 10:08:04
Document(s) archivé(s) le : jeudi 9 juin 2011 - 02:45:01

verbeek03asci.pdf

Fichiers produits par l'(les) auteur(s)

Non-linear feature extraction by the coordination of mixture models

Fichiers

Identifiants

Citation

Exporter

Métriques

226

526

Contact

Non-linear feature extraction by the coordination of mixture models

Fichiers

Identifiants

Citation

Exporter

Partager

Métriques

226

526

Contact