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

Switching between selection and fusion in combining classifiers: an experiment.

L I Kuncheva1

  • 1Sch. of Informatics, Univ. of Wales, Bangor.

IEEE Transactions on Systems, Man, and Cybernetics. Part B, Cybernetics : a Publication of the IEEE Systems, Man, and Cybernetics Society
|February 2, 2008
PubMed
Summary
This summary is machine-generated.

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

How good are fuzzy If-Then classifiers?

IEEE transactions on systems, man, and cybernetics. Part B, Cybernetics : a publication of the IEEE Systems, Man, and Cybernetics Society·2008
Same author

Presupervised and post-supervised prototype classifier design.

IEEE transactions on neural networks·2008
Same author

Fuzzy diagnosis.

Artificial intelligence in medicine·1999
Same author

Evaluation of computerized medical diagnostic decisions via fuzzy sets.

International journal of bio-medical computing·1991
Same author

Fuzzy multi-level classifier for medical applications.

Computers in biology and medicine·1990
Same journal

Strategic Ability Updating in Concurrent Games by Coalitional Commitment.

IEEE transactions on systems, man, and cybernetics. Part B, Cybernetics : a publication of the IEEE Systems, Man, and Cybernetics Society·2015
Same journal

Meta-Analysis of the First Facial Expression Recognition Challenge.

IEEE transactions on systems, man, and cybernetics. Part B, Cybernetics : a publication of the IEEE Systems, Man, and Cybernetics Society·2012
Same journal

Adjustable model-based fusion method for multispectral and panchromatic images.

IEEE transactions on systems, man, and cybernetics. Part B, Cybernetics : a publication of the IEEE Systems, Man, and Cybernetics Society·2012
Same journal

Face Feature Weighted Fusion Based on Fuzzy Membership Degree for Video Face Recognition.

IEEE transactions on systems, man, and cybernetics. Part B, Cybernetics : a publication of the IEEE Systems, Man, and Cybernetics Society·2012
Same journal

A New Adaptive Fast Cellular Automaton Neighborhood Detection and Rule Identification Algorithm.

IEEE transactions on systems, man, and cybernetics. Part B, Cybernetics : a publication of the IEEE Systems, Man, and Cybernetics Society·2012
Same journal

Human-arm-and-hand-dynamic model with variability analyses for a stylus-based haptic interface.

IEEE transactions on systems, man, and cybernetics. Part B, Cybernetics : a publication of the IEEE Systems, Man, and Cybernetics Society·2012
See all related articles

This study introduces a novel classifier combination strategy (CS+DT) that intelligently switches between classifier selection and fusion. This approach enhances performance by leveraging classifier strengths in specific feature spaces, outperforming individual methods and simpler combinations.

Area of Science:

  • Machine Learning
  • Artificial Intelligence
  • Pattern Recognition

Background:

  • Classifier combination techniques are crucial for improving model performance and robustness.
  • Existing methods like dynamic classifier selection (DCS) and simple fusion rules have limitations.
  • Intelligent switching between selection and fusion based on feature space characteristics is underexplored.

Purpose of the Study:

  • To propose and evaluate a novel hybrid classifier combination scheme, termed clustering-and-selection plus decision templates (CS+DT).
  • To compare the performance of CS+DT against its individual components and various established classifier combination methods.
  • To investigate the conditions under which classifier selection can be misled by ensemble diversity.

Main Methods:

Related Experiment Videos

  • A hybrid approach combining classifier selection (clustering-and-selection, CS) in dominant regions and classifier fusion (decision templates, DT) in other regions.
  • Experimental comparison using five datasets with both homogeneous (multilayer perceptrons) and heterogeneous classifier ensembles.
  • Evaluation against baseline methods including majority vote, naive Bayes, joint-distribution methods (BKS), DCS_LA, and simple fusion techniques (max, min, average, product).

    • Main Results:

    • The proposed CS+DT scheme demonstrated superior performance compared to its individual components (CS and DT) and most other evaluated methods.
    • The study identified scenarios where classifier selection, both static and dynamic, can be suboptimal due to inherent differences within the classifier ensemble.
    • The effectiveness of the CS+DT approach was validated across diverse datasets and ensemble configurations.

    Conclusions:

    • The CS+DT method offers an effective strategy for combining classifiers, balancing selection and fusion for improved predictive accuracy.
    • Understanding the interplay between ensemble diversity and classifier selection mechanisms is critical for designing robust combination strategies.
    • The proposed method provides a valuable alternative to existing classifier combination techniques, particularly in complex pattern recognition tasks.