Jove
Visualize
Contact Us
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

Related Experiment Videos

Validation of voting committees

E Bax1

  • 1Computer Science Department, California Institute of Technology, Pasadena, California 91125, USA.

Neural Computation
|June 6, 1998
PubMed
Summary
This summary is machine-generated.

This study introduces a method to estimate test errors for classifier voting committees. Validating individual classifiers is more effective than direct committee validation for bounding errors.

Related Concept Videos

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by
Same author

Partition-based and sharp uniform error bounds.

IEEE transactions on neural networksยท2008
See all related articles

Area of Science:

  • Machine Learning
  • Statistical Learning Theory

Background:

  • Directly validating all possible voting committees is computationally infeasible.
  • Existing methods struggle to provide useful error bounds for large committee pools.

Purpose of the Study:

  • To develop a novel method for bounding the test errors of voting committees.
  • To enable effective error estimation when the number of potential committees is vast.

Main Methods:

  • Proposing a validation strategy focusing on individual classifiers rather than entire committees.
  • Utilizing linear programming to infer committee error bounds from individual classifier performance.
  • Applying the method to real-world credit card transaction data.

Main Results:

Related Experiment Videos

  • Demonstrated the efficacy of the proposed method in bounding test errors.
  • Showcased the practical application using credit card data.
  • Extended the methodology for general classifier error bound inference.

Conclusions:

  • Individual classifier validation offers a more tractable approach to committee error bounding.
  • The developed method provides a practical solution for error estimation in large-scale classifier ensembles.
  • The approach is generalizable to inferring error bounds for individual classifiers.