Model Selection and Error Estimation

  • Bartlett P
  • Boucheron S
  • Lugosi G
N/ACitations
Citations of this article
21Readers
Mendeley users who have this article in their library.
Get full text

Abstract

We study model selection strategies based on penalized empirical loss minimization. We point out a tight relationship between error estimation and data-based complexity penalization: any good error estimate may be converted into a data-based penalty function and the performance of the estimate is governed by the quality of the error estimate. We consider several penalty functions, involving error estimates on independent test data, empirical {\sc vc} dimension, empirical {\sc vc} entropy, and margin-based quantities. We also consider the maximal difference between the error on the first half of the training data and the second half, and the expected maximal discrepancy, a closely related capacity estimate that can be calculated by Monte Carlo integration. Maximal discrepancy penalty functions are appealing for pattern classification problems, since their computation is equivalent to empirical risk minimization over the training data with some labels flipped.

Cite

CITATION STYLE

APA

Bartlett, P. L., Boucheron, S., & Lugosi, G. (2005). Model Selection and Error Estimation. SSRN Electronic Journal. https://doi.org/10.2139/ssrn.248567

Register to see more suggestions

Mendeley helps you to discover research relevant for your work.

Already have an account?

Save time finding and organizing research with Mendeley

Sign up for free