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

How good are support vector machines?

S Raudys1

  • 1Institute of Mathematics and Informatics, Vilnius, Lithuania. raudys@das.mii.lt

Neural Networks : the Official Journal of the International Neural Network Society
|August 10, 2000
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

On dimensionality, sample size, classification error, and complexity of classification algorithm in pattern recognition.

IEEE transactions on pattern analysis and machine intelligence·2011
Same author

k-nearest neighbors directed noise injection in multilayer perceptron training.

IEEE transactions on neural networks·2008
Same author

Evolution and generalization of a single neurone: I. Single-layer perceptron as seven statistical classifiers.

Neural networks : the official journal of the International Neural Network Society·2003
Same author

Evolution and generalization of a single neurone: II. Complexity of statistical classifiers and sample size considerations.

Neural networks : the official journal of the International Neural Network Society·2003
Same author

Evolution and generalization of a single neurone. III. Primitive, regularized, standard, robust and minimax regressions.

Neural networks : the official journal of the International Neural Network Society·2000
Same author

Visual classification of medical data using MLP mapping.

Computers in biology and medicine·1998
Same journal

TraNce: Type-aware hypergraph neural network with biological mediators for drug repositioning.

Neural networks : the official journal of the International Neural Network Society·2026
Same journal

Decentralized ADMM for factorization-based Low-rank matrix estimation.

Neural networks : the official journal of the International Neural Network Society·2026
Same journal

Memristive neuromorphic circuit design inspired by the neural mechanisms of conditioned fear.

Neural networks : the official journal of the International Neural Network Society·2026
Same journal

Q-learning based asynchronous Boolean control networks stabilization with data loss.

Neural networks : the official journal of the International Neural Network Society·2026
Same journal

New results on prescribed-time synchronization of complex networks via intermittent control.

Neural networks : the official journal of the International Neural Network Society·2026
Same journal

Variance-constrained multi-view ensemble broad network for imbalanced data.

Neural networks : the official journal of the International Neural Network Society·2026
See all related articles

Support vector (SV) machines may not be optimal for non-Gaussian data. A trained, optimally stopped single-layer perceptron (SLP) in a transformed feature space offers a better alternative for classification tasks.

Area of Science:

  • Machine Learning
  • Statistical Classification

Background:

  • Support vector (SV) machines are widely used for classification, particularly for data with abrupt density function changes.
  • However, SV classifiers may not be optimal for non-Gaussian data or data lacking sharp linear boundaries.
  • The performance of SV machines can be limited when numerous support vectors influence classification weights.

Purpose of the Study:

  • To evaluate the optimality of Support Vector (SV) classifiers for Gaussian data models.
  • To identify alternative classification methods that perform better on real-world, non-ideal data.
  • To explore the effectiveness of single-layer perceptrons (SLPs) in transformed feature spaces.

Main Methods:

  • Analysis of Support Vector (SV) classifier performance on a specific class of Gaussian data models.

Related Experiment Videos

  • Comparison of SV classifiers with other methods, considering the impact of multiple support vectors.
  • Implementation and evaluation of a specially trained and optimally stopped single-layer perceptron (SLP) in a transformed feature space.

    • Main Results:

    • The Support Vector (SV) classifier was found to be suboptimal for at least one class of Gaussian data based on mean generalization error.
    • SV classifiers can be outperformed by other methods when a large number of support vectors are involved.
    • A single-layer perceptron (SLP), when appropriately trained and stopped in a decorrelated and scaled feature space, presents a viable alternative.

    Conclusions:

    • Linear Support Vector (SV) machines are not universally optimal, especially for complex, real-world datasets.
    • Data transformation techniques, including decorrelation and scaling, are crucial for enhancing classifier performance.
    • Optimally stopped single-layer perceptrons (SLPs) in transformed feature spaces offer a promising alternative to linear SV machines for challenging classification problems.