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 vector quantization and the EM algorithm.

Ethem Alpaydin1

  • 1Department of Computer Engineering, Boğaziçi University, Istanbul, Turkey

Neural Networks : the Official Journal of the International Neural Network Society
|March 29, 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

Continuously Constructive Deep Neural Networks.

IEEE transactions on neural networks and learning systems·2019
Same author

Design and analysis of classifier learning experiments in bioinformatics: survey and case studies.

IEEE/ACM transactions on computational biology and bioinformatics·2012
Same author

Multiclass posterior probability support vector machines.

IEEE transactions on neural networks·2008
Same author

Ordering and finding the best of K > 2 supervised learning algorithms.

IEEE transactions on pattern analysis and machine intelligence·2006
Same author

ANNSyS: an Analog Neural Network Synthesis System.

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

Dynamic analysis and reliable mechanical optimization application of ring HNN effected with a memristive neuron.

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

DAFF-Net: A detection and search method for small-scale low surface brightness galaxies.

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

Quasi-synchronization for complex networks with hybrid pinning intermittent control.

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

This study reveals that training methods for fuzzy clustering models approximate the expectation-maximization (EM) algorithm for Gaussian mixtures. Probabilistic Gaussian mixtures offer advantages like interpretable variance and guaranteed convergence, outperforming other methods on the IRIS dataset.

Area of Science:

  • Machine Learning
  • Data Mining
  • Pattern Recognition

Background:

  • Existing fuzzy clustering algorithms like hard c-means (HCM), fuzzy c-means (FCM), fuzzy learning vector quantization (FLVQ), and soft competition scheme (SCS) share relations with probabilistic Gaussian mixtures (GM).
  • The training procedures of these models have been recently linked to approximations of the expectation-maximization (EM) algorithm used for Gaussian mixtures.

Purpose of the Study:

  • To extend the known relations between fuzzy clustering methods and Gaussian mixtures to their training procedures.
  • To demonstrate that learning rules in these models approximate the EM method for Gaussian mixtures.
  • To highlight the advantages of the probabilistic framework over traditional fuzzy clustering approaches.

Main Methods:

  • Comparative analysis of training algorithms for HCM, FCM, FLVQ, SCS, and GM.

Related Experiment Videos

  • Utilizing the expectation-maximization (EM) algorithm for training Gaussian mixtures.
  • Evaluating performance on the IRIS dataset, comparing EM-trained GM with LVQ (HCM), SCS, and FLVQ.

    • Main Results:

    • Fuzzy clustering training rules are shown to be approximations of the EM method for Gaussian mixtures.
    • Gaussian mixtures trained with EM demonstrate superior accuracy on the IRIS dataset due to covariance information utilization.
    • The probabilistic framework offers advantages including interpretable variance estimation, guaranteed convergence, and estimation of distance norms and cluster priors.

    Conclusions:

    • The expectation-maximization (EM) algorithm provides a robust probabilistic framework for clustering, offering significant advantages over traditional fuzzy methods.
    • A coupled approach to training cluster parameters and mappings is advocated for supervised learning, improving upon uncoupled methods.
    • Gaussian mixtures trained via EM represent a more accurate and interpretable clustering approach, especially when dealing with data exhibiting varying covariances.