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

A generalized discriminant rule when training population and test population differ on their descriptive parameters.

Christophe Biernacki1, Farid Beninel, Vincent Bretagnolle

  • 1Université de Besançon, UMR CNRS 6623, France. biernac math.univ-fcomte.fr

Biometrics
|June 20, 2002
PubMed
Summary

This study introduces a new discriminant analysis method for cases where labeled and unlabeled data come from slightly different populations. The approach uses linear mapping to improve classification accuracy in such scenarios.

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

<i>Territory health</i>: A transdisciplinary approach of environmental, plant, animal, and human health interdependencies in a rural landscape in France.

iScience·2026
Same author

First evidence of deleterious effect of pesticide mixture on health status in semi-captive grey partridges.

Environmental research·2025
Same author

Crop yield loss under high insecticide regime driven by reduction in natural pest control.

Proceedings. Biological sciences·2025
Same author

Pesticide mixture effects on physiological stress and morphology of growing wild nestlings.

Journal of hazardous materials·2025
Same author

Pesticide exposure in farmland wild passerines: bio-indicators of a widespread contamination despite organic farming.

Environmental research·2025
Same author

Economic efficiency of nature-based solutions: Theoretical framework and application to semi-natural habitat implementation.

Journal of environmental management·2024

Area of Science:

  • Statistics
  • Machine Learning
  • Bioinformatics

Background:

  • Standard discriminant analysis assumes identical populations for training and application.
  • Real-world data often exhibits slight population shifts, impacting classification accuracy.

Purpose of the Study:

  • To develop a robust discriminant analysis method for situations with differing populations.
  • To improve classification accuracy when training and application data are not from the same distribution.

Main Methods:

  • The study models the relationship between differing populations using linear mapping within a multinormal context.
  • Parameters of this linear relationship are estimated to adapt the discriminant rule.
  • Proposed models and estimation methods are presented.

Related Experiment Videos

Main Results:

  • The proposed method demonstrates improved performance compared to standard allocation rules in experimental illustrations.
  • An experimental case study involving morphometric sex determination in birds is presented.
  • The method's effectiveness is validated through comparison with traditional techniques.

Conclusions:

  • The developed method effectively handles population differences in discriminant analysis.
  • Linear mapping provides a viable approach for adapting discriminant rules to new populations.
  • The technique offers a valuable extension for partially labeled sample scenarios.