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

Soft learning vector quantization.

Sambu Seo1, Klaus Obermayer

  • 1Department of Electrical Engineering and Computer Science, Technical University of Berlin, 10587 Berlin, Germany. sontag@cs.tu-berlin.de

Neural Computation
|June 21, 2003
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

Polarity-dependent modulation of sleep oscillations and cortical excitability in aging.

Frontiers in aging neuroscience·2026
Same author

On the robustness of the emergent spatiotemporal dynamics in biophysically realistic and phenomenological whole-brain models at multiple network resolutions.

Frontiers in network physiology·2025
Same author

Neural dynamics underlying the cue validity effect in target conflict resolution.

Cerebral cortex (New York, N.Y. : 1991)·2025
Same author

A robotics-inspired scanpath model reveals the importance of uncertainty and semantic object cues for gaze guidance in dynamic scenes.

Journal of vision·2025
Same author

A framework for optimal control of oscillations and synchrony applied to non-linear models of neural population dynamics.

Frontiers in computational neuroscience·2024
Same author

Gaze Behavior Reveals Expectations of Potential Scene Changes.

Psychological science·2024
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

This study introduces novel Learning Vector Quantization (LVQ) algorithms derived from a principled Gaussian mixture model approach. These new methods improve classification performance and offer greater adaptability compared to existing LVQ 2.1 techniques.

Area of Science:

  • Machine Learning
  • Pattern Recognition
  • Statistical Modeling

Background:

  • Learning Vector Quantization (LVQ) is a widely used adaptive nearest prototype classifier for multiclass problems.
  • Existing LVQ algorithms are primarily based on heuristic principles, lacking a rigorous theoretical foundation.

Purpose of the Study:

  • To develop principled variants of LVQ algorithms using a Gaussian mixture model (GMM) ansatz.
  • To introduce a new objective function based on a likelihood ratio and derive a gradient descent learning rule.
  • To extend LVQ capabilities to accommodate different distance measures and enable "soft" classification.

Main Methods:

  • Derivation of two new LVQ variants grounded in a GMM framework.
  • Formulation of an objective function utilizing a likelihood ratio.

Related Experiment Videos

  • Application of gradient descent for learning rule derivation.

    • Main Results:

    • The proposed LVQ variants demonstrate superior classification performance compared to the established LVQ 2.1.
    • The new approach facilitates the integration of diverse distance metrics within LVQ.
    • The method allows for the creation of "soft" LVQ algorithms, offering probabilistic outputs.

    Conclusions:

    • The principled derivation of LVQ algorithms offers a more robust theoretical basis for adaptive nearest prototype classification.
    • The explicit model assumptions enhance the adaptability of the method to various problem types.
    • The developed LVQ variants represent a significant advancement in classification performance and flexibility.