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

Boosted mixture of experts: an ensemble learning scheme

R Avnimelech1, N Intrator

  • 1Department of Computer Science, Tel-Aviv University, Ramat-Aviv, 69978, Israel.

Neural Computation
|February 9, 1999
PubMed
Summary

This study introduces a novel supervised learning method for ensemble machines, combining predictor outputs using a dynamic model. This approach offers a new perspective on mixture of experts and boosting algorithms for classification tasks.

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

Competitive learning in biological and artificial neural computation.

Trends in cognitive sciences·2011
Same author

Multimodality exploration by an unsupervised projection pursuit neural network.

IEEE transactions on neural networks·2008
Same author

Targeting c-Myc-activated genes with a correlation method: detection of global changes in large gene expression network dynamics.

Proceedings of the National Academy of Sciences of the United States of America·2005
Same author

The effect of noise on a class of energy-based learning rules.

Neural computation·2003
Same author

(Coarse coding of shape fragments) + (retinotopy) approximately = representation of structure.

Spatial vision·2001
Same author

Solutions of the BCM learning rule in a network of lateral interacting nonlinear neurons.

Network (Bristol, England)·1999

Area of Science:

  • Machine Learning
  • Computer Science

Background:

  • Ensemble methods are crucial for improving predictive accuracy in machine learning.
  • Existing methods like mixture of experts and boosting algorithms have limitations in dynamic output combination.

Purpose of the Study:

  • To introduce a new supervised learning procedure for ensemble machines.
  • To present a dynamic classifier combination model for integrating predictor outputs.
  • To explore the relationship of this method with mixture of experts and boosting algorithms.

Main Methods:

  • A supervised learning procedure is proposed for ensemble machines.
  • Outputs from predictors trained on different distributions are combined using a dynamic classifier combination model.
  • The method is framed as a variant of mixture of experts or boosting algorithms.

Related Experiment Videos

Main Results:

  • The proposed procedure can be adapted for general classification and regression problems.
  • It offers advantages over classical ensemble approaches, particularly in dynamic output combination.
  • Demonstrated effectiveness on a synthetic dataset and a NIST digit recognition task.

Conclusions:

  • The novel supervised learning procedure provides an effective way to combine predictor outputs in ensemble machines.
  • This method enhances existing mixture of experts and boosting techniques.
  • It shows promise for improving performance in classification and regression tasks.