Cross-validation improved by aggregation: Agghoo - Inria - Institut national de recherche en sciences et technologies du numérique Accéder directement au contenu
Pré-Publication, Document De Travail Année : 2017

Cross-validation improved by aggregation: Agghoo

Résumé

Cross-validation is widely used for selecting among a family of learning rules. This paper studies a related method, called aggregated hold-out (Agghoo), which mixes cross-validation with aggregation; Agghoo can also be related to bagging. According to numerical experiments, Agghoo can improve significantly cross-validation's prediction error, at the same computational cost; this makes it very promising as a general-purpose tool for prediction. We provide the first theoretical guarantees on Agghoo, in the supervised classification setting, ensuring that one can use it safely: at worse, Agghoo performs like the hold-out, up to a constant factor. We also prove a non-asymptotic oracle inequality, in binary classification under the margin condition, which is sharp enough to get (fast) minimax rates.
Fichier principal
Vignette du fichier
agregcv_nips2017.pdf (310.44 Ko) Télécharger le fichier
Origine : Fichiers produits par l'(les) auteur(s)
Loading...

Dates et versions

hal-01585595 , version 1 (11-09-2017)

Identifiants

Citer

Guillaume Maillard, Sylvain Arlot, Matthieu Lerasle. Cross-validation improved by aggregation: Agghoo. 2017. ⟨hal-01585595⟩
493 Consultations
256 Téléchargements

Altmetric

Partager

Gmail Facebook X LinkedIn More