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

Three learning phases for radial-basis-function networks.

F Schwenker1, H A Kestler, G Palm

  • 1Department of Neural Information Processing, University of Ulm, Germany. schwenker@informatik.uni-ulm.de

Neural Networks : the Official Journal of the International Neural Network Society
|June 20, 2001
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

Bringing Anatomical Information into Neuronal Network Models.

Advances in experimental medicine and biology·2022
Same author

[Tumor biology of oropharyngeal carcinoma].

HNO·2020
Same author

Detection of herpesvirus capsids in transmission electron microscopy images using transfer learning.

Histochemistry and cell biology·2018
Same author

Erratum: "Molecular radiotherapy: The NUKFIT software for calculating the time-integrated activity coefficient" [Med. Phys. 40, 102504 (2013)].

Medical physics·2017
Same author

RUNX1 mutations in acute myeloid leukemia are associated with distinct clinico-pathologic and genetic features.

Leukemia·2016
Same author

RUNX1 mutations in acute myeloid leukemia are associated with distinct clinico-pathologic and genetic features.

Leukemia·2016
Same journal

Exploiting audio-visual modalities in videos: Object detection via multi-stage bilateral coupling network.

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

Reliability-aware modality completion with cross-modal distillation for federated learning with missing modalities.

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

IGFD-Net: Illumination-guided frequency decoupling for polarization image fusion.

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

Multiple-Strategies dung beetle optimizer and its applications in engineering optimization and bankruptcy prediction.

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

Aggregating global-scale pixel-wise forgery cues within a graph.

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

Finite-Time intermittent control for secure synchronization of Neutral-Type stochastic delayed neural networks under aperiodic DoS attacks.

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

This study explores radial basis function (RBF) network learning algorithms, comparing two-phase, three-phase, and support vector (SV) learning. Three-phase and SV learning show superior performance over two-phase learning in pattern recognition tasks.

Area of Science:

  • Machine Learning
  • Artificial Intelligence
  • Pattern Recognition

Background:

  • Radial Basis Function (RBF) networks offer alternative learning schemes compared to Multilayer Perceptrons (MLPs).
  • RBF network training can be categorized into one-, two-, and three-phase learning strategies.
  • Two-phase learning is common, training RBF layers and output layers separately.

Purpose of the Study:

  • To investigate and compare different learning algorithms for Radial Basis Function (RBF) networks.
  • To evaluate the effectiveness of two-phase, three-phase, and Support Vector (SV) learning schemes.
  • To demonstrate performance improvements using three-phase learning and analyze SV learning characteristics.

Main Methods:

  • Categorization of RBF learning into one-, two-, and three-phase schemes.

Related Experiment Videos

  • Implementation of two-phase learning: separate training of RBF centers/scaling and output weights.
  • Introduction of three-phase learning: simultaneous adaptation of all RBF network parameters.
  • Application of Support Vector (SV) learning as a specialized one-phase approach.

    • Main Results:

    • Numerical experiments were conducted on three distinct pattern recognition applications: 3D object classification, handwritten digit recognition, and electrocardiogram categorization.
    • Three-phase learning demonstrated performance improvements over standard two-phase learning by adapting all parameters simultaneously.
    • Support Vector (SV) learning and three-phase learning outperformed two-phase learning, though SV learning can result in complex network structures.

    Conclusions:

    • Three-phase learning offers a practical enhancement to RBF network performance, particularly when combined with unlabeled data.
    • Support Vector (SV) learning provides a competitive alternative but may lead to increased network complexity.
    • The choice of learning scheme impacts RBF network performance and structural complexity in pattern recognition tasks.