Computing the nonnegative 3-way tensor factorization using Tikhonov regularization

Jean-Philip Royer 1 Pierre Comon 1 Nadège Thirion 2
1 Laboratoire d'Informatique, Signaux, et Systèmes de Sophia-Antipolis (I3S) / Equipe SIGNAL
SIS - Signal, Images et Systèmes
2 LSEET - Laboratoire de sondages électromagnétiques de l'environnement terrestre
Abstract : This paper deals with the minimum polyadic decomposition of a nonnegative three-way array. The main advantage of the nonnegativity constraint is that the approximation problem becomes well posed. To tackle this problem, we suggest the use of a cost function including penalty terms built with matrix exponentials. Gradient components are then derived, allowing to efficiently implement the decomposition using classical optimization algorithms. In our case, Alternating Least Squares (ALS) and conjugate gradient algorithms are studied and compared with another existing algorithm, thanks to computer simulations performed in the context of data analysis.
Keywords : Data analysis data mining non linear conjugate gradient nonnegative 3-way array Canonical Polyadic decomposition Tikhonov regularization
Type de document :
Communication dans un congrès
IEEE. ICASSP, May 2011, Prague, Czech Republic. IEEE, pp.2732-2735, 2011
Liste complète des métadonnées

Littérature citée [11 références]  Voir  Masquer  Télécharger
https://hal.archives-ouvertes.fr/hal-00641065
Contributeur : Pierre Comon <>
Soumis le : lundi 14 novembre 2011 - 17:04:01
Dernière modification le : vendredi 7 mars 2014 - 08:42:46
Document(s) archivé(s) le : mercredi 15 février 2012 - 02:32:24

Fichier

hal-Tikhonov.pdf
Fichiers produits par l'(les) auteur(s)

Identifiants

  • HAL Id : hal-00641065, version 1

Collections

UNICE | INSU | I3S | OSU-INSTITUT-PYTHEAS | LSEET

Citation

Jean-Philip Royer, Pierre Comon, Nadège Thirion. Computing the nonnegative 3-way tensor factorization using Tikhonov regularization. IEEE. ICASSP, May 2011, Prague, Czech Republic. IEEE, pp.2732-2735, 2011. 〈hal-00641065〉

Exporter

BibTeX TEI DC DCterms EndNote

Partager

Métriques

Consultations de
la notice

348

Téléchargements du document

125

Contact

support.ccsd.cnrs.fr
hal.support@ccsd.cnrs.fr
Logo CCSD