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

Boosting neural networks.

H Schwenk1, Y Bengio

  • 1LIMSI-CNRS, 91403 Orsay cedex, France.

Neural Computation
|August 23, 2000
PubMed
Summary

AdaBoost (Adaptive Boosting) effectively enhances neural network performance, outperforming boosted decision trees. Random resampling is not the key to AdaBoost

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

Adaptive importance sampling to accelerate training of a neural probabilistic language model.

IEEE transactions on neural networks·2008
Same author

Taking on the curse of dimensionality in joint distributions using neural networks.

IEEE transactions on neural networks·2008
Same author

Cost functions and model combination for VaR-based asset allocation using neural networks.

IEEE transactions on neural networks·2008
Same author

Experiments on the application of IOHMMs to model financial returns series.

IEEE transactions on neural networks·2008
Same author

Bias learning, knowledge sharing.

IEEE transactions on neural networks·2008
Same author

Locally linear embedding for dimensionality reduction in QSAR.

Journal of computer-aided molecular design·2005

Area of Science:

  • Machine Learning
  • Artificial Intelligence
  • Computer Science

Background:

  • Boosting algorithms enhance learning model performance.
  • AdaBoost (Adaptive Boosting) is a successful boosting algorithm, often using decision trees.
  • The efficacy of AdaBoost with neural networks requires investigation.

Purpose of the Study:

  • To evaluate AdaBoost's performance with neural networks.
  • To compare different AdaBoost training methods (sampling vs. cost function weighting).
  • To understand the role of random resampling in AdaBoost's success.

Main Methods:

  • Investigated AdaBoost with multilayer neural networks.
  • Compared training methods: random resampling and cost function weighting.
  • Evaluated performance on benchmark datasets (handwritten digits, UCI letters, UCI satellite).

Main Results:

  • AdaBoost with neural networks achieved competitive error rates (1.4% on handwritten digits, 1.5% on UCI letters, 8.1% on UCI satellite).
  • Random resampling of training data is not the primary driver of AdaBoost's performance gains.
  • AdaBoost with neural networks outperformed boosted decision trees on tested datasets.

Conclusions:

  • AdaBoost is effective when combined with neural networks, offering significant performance improvements.
  • The success of AdaBoost is attributed to mechanisms beyond simple random data resampling.
  • AdaBoost provides a valuable alternative to traditional boosting methods using decision trees.

Related Experiment Videos