4455 articles – 13149 Notices  [english version]
HAL : hal-00725289, version 1

Fiche détaillée  Récupérer au format
20th European Signal Processing Conference (EUSIPCO-2012), Bucarest : Romania (2012)
NonNegative 3-Way tensor Factorzation taking into account Possible Missing Data
Jean-Philip Royer 1, 2, Nadège Thirion-Moreau 2, Pierre Comon ( ) 3
(26/08/2012)

The paper deals with the problem of incomplete data i.e. data with missing, unknown or unreliable values, in the polyadic decomposition of a nonnegative three-way tensor. The main advantage of the nonnegativity constraint is that the approximation problem becomes well posed. To tackle simultaneously these two problems, we suggest the use of a weighted least square cost function whose weights are gradually modified through the iterations. Moreover, the nonnegative nature of the loading matrices is taken into account directly in the problem parameterization. Then, the three gradient components can be explicitly derived allowing to efficiently implement the CP decomposition using standard optimization algorithms. In our case, we focus on the conjugate gradient and the BFGS algorithms. Finally, the good behavior of the proposed approaches and their robustness versus possible model errors is illustrated through computer simulations in the context of data analysis.
1 :  Laboratoire d'Informatique, Signaux, et Systèmes de Sophia-Antipolis (I3S) / Equipe SIGNAL
Université Nice Sophia Antipolis [UNS] – CNRS : UMR7271
2 :  Laboratoire des Sciences de l'Information et des Systèmes (LSIS)
CNRS : UMR7296 – Arts et Métiers ParisTech – Université Paul Cézanne - Aix-Marseille III – Université de la Méditerranée - Aix-Marseille II – Université de Provence - Aix-Marseille I – Université Sud Toulon Var
3 :  Grenoble Images Parole Signal Automatique (GIPSA-lab)
CNRS : UMR5216 – Université Joseph Fourier - Grenoble I – Université Pierre-Mendès-France - Grenoble II – Université Stendhal - Grenoble III – Institut Polytechnique de Grenoble - Grenoble Institute of Technology
SIGNAL
CICS
Informatique/Traitement du signal et de l'image

Sciences de l'ingénieur/Traitement du signal et de l'image
Data analysis – Conjugate gradient – Canonical – Polyadic – Decomposition – Tensor – CanDecomp – Parafac – Fluorescence – Spectrometry
Liste des fichiers attachés à ce document : 
PDF
RoyeTC12-Bucarest.pdf(572.1 KB)