Access

You are not currently logged in.

Access your personal account or get JSTOR access through your library or other institution:

login

Log in to your personal account or through your institution.

Linear Model Selection by Cross-Validation

Jun Shao
Journal of the American Statistical Association
Vol. 88, No. 422 (Jun., 1993), pp. 486-494
DOI: 10.2307/2290328
Stable URL: http://www.jstor.org/stable/2290328
Page Count: 9
  • Download ($14.00)
  • Cite this Item
Linear Model Selection by Cross-Validation
Preview not available

Abstract

We consider the problem of selecting a model having the best predictive ability among a class of linear models. The popular leave-one-out cross-validation method, which is asymptotically equivalent to many other model selection methods such as the Akaike information criterion (AIC), the Cp, and the bootstrap, is asymptotically inconsistent in the sense that the probability of selecting the model with the best predictive ability does not converge to 1 as the total number of observations n → ∞. We show that the inconsistency of the leave-one-out cross-validation can be rectified by using a leave-nν-out cross-validation with nν, the number of observations reserved for validation, satisfying nν/n → 1 as n → ∞. This is a somewhat shocking discovery, because nν/n → 1 is totally opposite to the popular leave-one-out recipe in cross-validation. Motivations, justifications, and discussions of some practical aspects of the use of the leave-nν-out cross-validation method are provided, and results from a simulation study are presented.

Page Thumbnails

  • Thumbnail: Page 
486
    486
  • Thumbnail: Page 
487
    487
  • Thumbnail: Page 
488
    488
  • Thumbnail: Page 
489
    489
  • Thumbnail: Page 
490
    490
  • Thumbnail: Page 
491
    491
  • Thumbnail: Page 
492
    492
  • Thumbnail: Page 
493
    493
  • Thumbnail: Page 
494
    494