21808 articles – 15604 references  [version française]

hal-00407906, version 1

A survey of cross-validation procedures for model selection

Sylvain Arlot () 1, Alain Celisse () 2

Statistics Surveys 4 (2010) 40--79

Abstract: Used to estimate the risk of an estimator or to perform model selection, cross-validation is a widespread strategy because of its simplicity and its apparent universality. Many results exist on the model selection performances of cross-validation procedures. This survey intends to relate these results to the most recent advances of model selection theory, with a particular emphasis on distinguishing empirical statements from rigorous theoretical results. As a conclusion, guidelines are provided for choosing the best cross-validation procedure according to the particular features of the problem in hand.

  • 1:  Laboratoire d'informatique de l'école normale supérieure (LIENS)
  • CNRS : UMR8548 – Ecole normale supérieure de Paris - ENS Paris
  • 2:  Mathématiques et Informatique Appliquées (MIA)
  • Institut national de la recherche agronomique (INRA) : UMR0518 – AgroParisTech
  • Collaboration : ANR-09-JCJC-0027-01
  • Domain : Mathematics/Statistics
    Statistics/Other Statistics
    Statistics/Statistics Theory
    Statistics/Applications
    Statistics/Methodology
  • Keywords : cross-validation – leave-one-out – model selection
  • Comment : Published in Statistics Surveys (2010) 4 – 40-79
 
  • hal-00407906, version 1
  • http://hal.archives-ouvertes.fr/hal-00407906
  • oai:hal.archives-ouvertes.fr:hal-00407906
  • From: Alain Celisse <>
  • Submitted on: Monday, 27 July 2009 19:24:11
  • Updated on: Thursday, 28 June 2012 10:17:03