VAITER, Samuel; DELEDALLE, Charles; PEYRÉ, Gabriel; FADILI, Jalal M.; DOSSAL, Charles

Métadonnées

Afficher la notice complète

Licence d’utilisation du document

VAITER, Samuel
CEntre de REcherches en MAthématiques de la DEcision [CEREMADE]

DELEDALLE, Charles
CEntre de REcherches en MAthématiques de la DEcision [CEREMADE]

PEYRÉ, Gabriel
CEntre de REcherches en MAthématiques de la DEcision [CEREMADE]

Langue

Document de travail - Pré-publication

Résumé en anglais

In this paper, we are concerned with regression problems where covariates can be grouped in nonoverlapping blocks, and where only a few of them are assumed to be active. In such a situation, the group Lasso is an at- tractive method for variable selection since it promotes sparsity of the groups. We study the sensitivity of any group Lasso solution to the observations and provide its precise local parameterization. When the noise is Gaussian, this allows us to derive an unbiased estimator of the degrees of freedom of the group Lasso. This result holds true for any fixed design, no matter whether it is under- or overdetermined. With these results at hand, various model selec- tion criteria, such as the Stein Unbiased Risk Estimator (SURE), are readily available which can provide an objectively guided choice of the optimal group Lasso fit.< Réduire

Mots clés en anglais

Group Lasso

Degrees of freedom

Sparsity

Model selection criteria

URI

https://oskar-bordeaux.fr/handle/20.500.12278/189736

Projet Européen

Sparsity, Image and Geometry to Model Adaptively Visual Processings

Origine

Importé de hal

Unités de recherche

Institut de Mathématiques de Bordeaux (IMB) - UMR 5251