An algebraic solution for the Candecomp/PARAFAC decomposition with circulant factors - Archive ouverte HAL Accéder directement au contenu
Article Dans Une Revue SIAM Journal on Matrix Analysis and Applications Année : 2014

An algebraic solution for the Candecomp/PARAFAC decomposition with circulant factors

Résumé

The Candecomp/PARAFAC decomposition (CPD) is an important mathematical tool used in several fields of application. Yet, its computation is usually performed with iterative methods which are subject to reaching local minima and to exhibiting slow convergence. In some practical contexts, the data tensors of interest admit decompositions constituted by matrix factors with particular structure. Often, such structure can be exploited for devising specialized algorithms with superior properties in comparison with general iterative methods. In this paper, we propose a novel approach for computing a circulant-constrained CPD (CCPD), i.e., a CPD of a hypercubic tensor whose factors are all circulant (and possibly tall). To this end, we exploit the algebraic structure of such tensor, showing that the elements of its frequency-domain counterpart satisfy homogeneous monomial equations in the eigenvalues of square circulant matrices associated with its factors, which we can therefore estimate by solving these equations. Then, we characterize the sets of solutions admitted by such equations under Kruskal's uniqueness condition. Simulation results are presented, validating our approach and showing that it can help avoiding typical disadvantages of iterative methods.
Fichier principal
Vignette du fichier
goulart-favier-CCPD_HAL.pdf (320.45 Ko) Télécharger le fichier
Origine : Fichiers produits par l'(les) auteur(s)
Loading...

Dates et versions

hal-00967263 , version 1 (28-03-2014)
hal-00967263 , version 2 (17-03-2015)

Identifiants

  • HAL Id : hal-00967263 , version 2

Citer

José Henrique de Morais Goulart, Gérard Favier. An algebraic solution for the Candecomp/PARAFAC decomposition with circulant factors. SIAM Journal on Matrix Analysis and Applications, 2014, 35 (4), pp.1543-1562. ⟨hal-00967263v2⟩
236 Consultations
732 Téléchargements

Partager

Gmail Facebook X LinkedIn More