Bayesian Nested Group Lasso

Marco Stefanucci; Pierfrancesco Alaimo Di Loro; Rosario Barone

doi:10.1007/978-3-031-96033-8_60

Bayesian Nested Group Lasso

Marco Stefanucci
, Pierfrancesco Alaimo Di Loro
, Rosario Barone

Risultato della ricerca: Contributo in libro › Contributo a conferenza

Abstract

We introduce the Bayesian Nested Group Lasso, a hierarchical model extending the Group Lasso to nested structures. By formulating the penalty as a scale mixture of Gaussians, we derive an efficient MCMC sampling scheme. A simulation study shows that our method reduces posterior variability for irrelevant variables, outperforming the Bayesian Lasso in structured sparsity settings. We discuss potential extensions, including spike-and-slab priors, for exact variable selection.

Lingua originale	Inglese
Titolo della pubblicazione ospite	Statistics for Innovation IV
Editore	Springer
Pagine	369-374
Numero di pagine	6
ISBN (stampa)	978-3-031-96035-2
DOI	https://doi.org/10.1007/978-3-031-96033-8_60
Stato di pubblicazione	Pubblicato - 2025

Keywords

Bayesian Lasso
Nested Group Lasso
Overlap Group Lasso

Accesso al documento

10.1007/978-3-031-96033-8_60

Altri file e collegamenti

Link a IRIS PubliCatt

Cita questo

@inproceedings{72dedd0912be4ec6a083084131c9a212,

title = "Bayesian Nested Group Lasso",

abstract = "We introduce the Bayesian Nested Group Lasso, a hierarchical model extending the Group Lasso to nested structures. By formulating the penalty as a scale mixture of Gaussians, we derive an efficient MCMC sampling scheme. A simulation study shows that our method reduces posterior variability for irrelevant variables, outperforming the Bayesian Lasso in structured sparsity settings. We discuss potential extensions, including spike-and-slab priors, for exact variable selection.",

keywords = "Bayesian Lasso, Nested Group Lasso, Overlap Group Lasso, Bayesian Lasso, Nested Group Lasso, Overlap Group Lasso",

author = "Marco Stefanucci and Loro, \{Pierfrancesco Alaimo Di\} and Rosario Barone",

year = "2025",

doi = "10.1007/978-3-031-96033-8\_60",

language = "English",

isbn = "978-3-031-96035-2",

pages = "369--374",

booktitle = "Statistics for Innovation IV",

publisher = "Springer",

}

TY - GEN

T1 - Bayesian Nested Group Lasso

AU - Stefanucci, Marco

AU - Loro, Pierfrancesco Alaimo Di

AU - Barone, Rosario

PY - 2025

Y1 - 2025

N2 - We introduce the Bayesian Nested Group Lasso, a hierarchical model extending the Group Lasso to nested structures. By formulating the penalty as a scale mixture of Gaussians, we derive an efficient MCMC sampling scheme. A simulation study shows that our method reduces posterior variability for irrelevant variables, outperforming the Bayesian Lasso in structured sparsity settings. We discuss potential extensions, including spike-and-slab priors, for exact variable selection.

AB - We introduce the Bayesian Nested Group Lasso, a hierarchical model extending the Group Lasso to nested structures. By formulating the penalty as a scale mixture of Gaussians, we derive an efficient MCMC sampling scheme. A simulation study shows that our method reduces posterior variability for irrelevant variables, outperforming the Bayesian Lasso in structured sparsity settings. We discuss potential extensions, including spike-and-slab priors, for exact variable selection.

KW - Bayesian Lasso

KW - Nested Group Lasso

KW - Overlap Group Lasso

KW - Bayesian Lasso

KW - Nested Group Lasso

KW - Overlap Group Lasso

UR - https://publicatt.unicatt.it/handle/10807/323989

U2 - 10.1007/978-3-031-96033-8_60

DO - 10.1007/978-3-031-96033-8_60

M3 - Conference contribution

SN - 978-3-031-96035-2

SP - 369

EP - 374

BT - Statistics for Innovation IV

PB - Springer

ER -

Bayesian Nested Group Lasso

Abstract

Keywords

Accesso al documento

Altri file e collegamenti

Fingerprint

Cita questo