A Contrast for Independent Component Analysis with Priors on Source Kurtosis Signs

Abstract : A contrast function for Independent Component Analysis (ICA) is presented incorporating the prior knowledge on the sub-Gaussian or super-Gaussian character of the sources as described by their kurtosis signs. The contrast is related to the maximum likelihood principle, reduces the permutation indeterminacy typical of ICA, and proves particularly useful in the direct extraction of a source signal with distinct kurtosis sign. In addition, its numerical maximization can be performed cost-effectively by a Jacobi-like pairwise iteration. Extensions to standardized cumulants of orders other than four are also given.
Type de document :
Article dans une revue
IEEE Signal Processing Letters, Institute of Electrical and Electronics Engineers, 2008, 15, pp.501-504. <10.1109/LSP.2008.919845>
Liste complète des métadonnées

https://hal.archives-ouvertes.fr/hal-00342353
Contributeur : Pierre Comon <>
Soumis le : jeudi 27 novembre 2008 - 11:29:00
Dernière modification le : mercredi 21 janvier 2009 - 16:16:43

Identifiants

Collections

Citation

Vicente Zarzoso, Ronald Phlypo, Pierre Comon. A Contrast for Independent Component Analysis with Priors on Source Kurtosis Signs. IEEE Signal Processing Letters, Institute of Electrical and Electronics Engineers, 2008, 15, pp.501-504. <10.1109/LSP.2008.919845>. <hal-00342353>

Partager

Métriques

Consultations de la notice

240