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

An Equivalence Between Sparse Approximation and Support Vector Machines

Girosi1

  • 1Massachusetts Institute of Technology, Center for Biological and Computational Learning, Cambridge MA, US, 02139. girosi@ai.mit.edu

Neural Computation
|August 11, 1998
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

A Sparse Representation for Function Approximation.

Neural computation·1998
Same journal

A Model-Free Reinforcement Learning Implementation of Decision Making Under Uncertainty by Sequential Sampling.

Neural computation·2026
Same journal

DROP: Distributional and Regular Optimism and Pessimism for Reinforcement Learning.

Neural computation·2026
Same journal

Hierarchical Active Inference Using Successor Representations.

Neural computation·2026
Same journal

W-Kernel and Its Principal Space for Frequentist Evaluation of Bayesian Estimators.

Neural computation·2026
Same journal

A Hidden Markov Model-Inspired Sequence Classification Method for Hyperdimensional Computing.

Neural computation·2026
Same journal

Sparse Graphical Modeling for Electrophysiological Phase-Based Connectivity Using Circular Statistics.

Neural computation·2026
See all related articles

Support vector machines (SVM) and basis pursuit denoising are linked. Under noiseless conditions, these approximation techniques yield identical solutions by solving the same quadratic programming problem.

Area of Science:

  • Machine Learning
  • Applied Mathematics
  • Signal Processing

Background:

  • Support Vector Machines (SVM) is a powerful machine learning technique derived from the structural risk minimization principle.
  • Sparse approximation, exemplified by basis pursuit denoising, reconstructs functions using minimal basis functions from a large dictionary.
  • Both SVM and basis pursuit denoising are approximation techniques with applications in various fields.

Purpose of the Study:

  • To establish a relationship between Support Vector Machines (SVM) and a sparse approximation technique similar to basis pursuit denoising.
  • To demonstrate the equivalence of these two methods under specific conditions.
  • To connect SVM, sparse approximation, and regularization theory.

Main Methods:

  • The study analyzes the mathematical formulations of SVM and basis pursuit denoising.

Related Experiment Videos

  • A modified basis pursuit denoising algorithm is proposed and compared to SVM.
  • The equivalence is shown by demonstrating that both techniques solve the same quadratic programming problem when applied to noiseless data.

    • Main Results:

    • A direct relationship is established between SVM and a modified basis pursuit denoising algorithm.
    • Under noiseless conditions, SVM and the modified basis pursuit denoising yield identical solutions.
    • The equivalence is mathematically proven through the solution of a shared quadratic programming problem.

    Conclusions:

    • Support Vector Machines (SVM) and basis pursuit denoising are fundamentally related approximation techniques.
    • The findings provide a unified perspective on these methods within regularization theory.
    • This research bridges concepts from statistical learning theory and sparse approximation.