On Cross Validation for Model Selection

Neural Comp., Vol. 11, No. 4. (15 May 1999), pp. 863-870.

View FullText article

HighWire, Pubmed, Hubmed

Reviews [Write a review of this article]

There are no reviews of this article

Notes for this article

haiyeong has 0 private notes и ещё 1 public note for this article.

By comparison, a statistical test is based on unbiased estimations of the noise variance s2 through the residuals of both models when (14) holds.

haiyeong (public ) - 2008-04-28 16:13:05

Find related articles from these CiteULike users

haiyeong

Find related articles with these CiteULike tags

cross_validation, model_selection, pattern_recognition

Abstract

In response to Zhu and Rower (1996), a recent communication (Goutte, 1997) established that leave-one-out cross validation is not subject to the "no-free-lunch" criticism. Despite this optimistic conclusion, we show here that cross validation has very poor performances for the selection of linear models as compared to classic statistical tests. We conclude that the statistical tests are preferable to cross validation for linear as well as for nonlinear model selection.

@article{citeulike:2730199, abstract = {In response to Zhu and Rower (1996), a recent communication (Goutte, 1997) established that leave-one-out cross validation is not subject to the "no-free-lunch" criticism. Despite this optimistic conclusion, we show here that cross validation has very poor performances for the selection of linear models as compared to classic statistical tests. We conclude that the statistical tests are preferable to cross validation for linear as well as for nonlinear model selection.}, author = {Rivals, Isabelle and Personnaz, Leon }, citeulike-article-id = {2730199}, comment = {By comparison, a statistical test is based on unbiased estimations of the noise variance s2 through the residuals of both models when (14) holds.}, journal = {Neural Comp.}, keywords = {cross\_validation, model\_selection, pattern\_recognition}, month = {May}, number = {4}, pages = {863--870}, posted-at = {2008-04-28 15:53:04}, priority = {0}, title = {On Cross Validation for Model Selection}, url = {http://neco.mitpress.org/cgi/content/abstract/11/4/863}, volume = {11}, year = {1999} }

RIS record

TY - JOUR ID - citeulike:2730199 L3 - citeulike-article-id:2730199 TI - On Cross Validation for Model Selection JF - Neural Comp. VL - 11 IS - 4 SP - 863 EP - 870 N2 - In response to Zhu and Rower (1996), a recent communication (Goutte, 1997) established that leave-one-out cross validation is not subject to the "no-free-lunch" criticism. Despite this optimistic conclusion, we show here that cross validation has very poor performances for the selection of linear models as compared to classic statistical tests. We conclude that the statistical tests are preferable to cross validation for linear as well as for nonlinear model selection. KW - cross_validation KW - model_selection KW - pattern_recognition N1 - By comparison, a statistical test is based on unbiased estimations of the noise variance s2 through the residuals of both models when (14) holds. AU - Rivals, Isabelle AU - Personnaz, Leon PY - 1999/05/15/ UR - http://neco.mitpress.org/cgi/content/abstract/11/4/863 ER -

RIS BibTeX