PAC-Bayesian aggregation of linear estimators - Inria - Institut national de recherche en sciences et technologies du numérique Access content directly
Preprints, Working Papers, ... Year : 2014

PAC-Bayesian aggregation of linear estimators

Abstract

We consider the aggregation of linear estimator in regresison with a sub-Gaussian noise assumption. Aggregating estimators using exponential weights depending on their risk performs well in expectation, but sadly not in probability. A way to overcome this issue is considering exponential weights of a penalized risk. In this case, an oracle inequality can be obtained in probability, but is not sharp. Taking into account the estimated function's norm in the penalty offers a sharp inequality.
Fichier principal
Vignette du fichier
pacewa.pdf (235.43 Ko) Télécharger le fichier
Origin : Files produced by the author(s)

Dates and versions

hal-01070805 , version 1 (02-10-2014)
hal-01070805 , version 2 (24-10-2016)
hal-01070805 , version 3 (30-01-2018)

Identifiers

Cite

Lucie Montuelle, Erwan Le Pennec. PAC-Bayesian aggregation of linear estimators. 2014. ⟨hal-01070805v1⟩

Collections

LM-ORSAY
621 View
509 Download

Altmetric

Share

Gmail Facebook X LinkedIn More