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

A Structure-Adaptive Hybrid RBF-BP Classifier with an Optimized Learning Strategy.

Hui Wen1, Weixin Xie1, Jihong Pei1

  • 1ATR Key Lab of National Defense, shenzhen University, shenzhen 518060, China.

Plos One
|October 30, 2016
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

Design, synthesis, and characterization of species-independent h/mGPR35 agonists for MAFLD/MASH.

Bioorganic chemistry·2026
Same author

Janus barrier films of shellac/chitosan with natamycin for photoprotective on-demand antifungal release and starch metabolism regulation in kiwifruit preservation.

Food research international (Ottawa, Ont.)·2026
Same author

The GIGANTEA-LHY complex regulates citrus cold tolerance and participates in low-temperature-induced flowering.

Plant physiology·2026
Same author

Noninvasive tracking of myocardial fibrosis in pressure-overload heart failure with [<sup>68</sup>Ga]Ga-FAPI-04 PET/CT.

EJNMMI research·2026
Same author

Chlorogenic acid modulates configuration-dependent structural organization, lipid oxidation, and digestion in aqueous starch-unsaturated fatty acid systems.

Food research international (Ottawa, Ont.)·2026
Same author

Inflammatory bowel disease and renal disorders: from clinical associations to shared mechanisms and management strategies.

Frontiers in immunology·2026

This study introduces a novel structure-adaptive hybrid RBF-BP (SAHRBF-BP) classifier. This advanced classifier demonstrates superior performance, especially on low-dimensional, large datasets, outperforming other shallow learning networks.

Area of Science:

  • Machine Learning
  • Artificial Intelligence
  • Computational Science

Background:

  • Traditional Radial Basis Function (RBF) and Backpropagation (BP) networks have limitations in adaptively handling complex data distributions.
  • Optimizing the number of hidden nodes and their parameters in RBF networks is crucial for effective nonlinear mapping and classification.
  • Existing shallow learning networks (SLFNs) often struggle with performance on low-dimensional, large-scale datasets.

Purpose of the Study:

  • To present a novel structure-adaptive hybrid RBF-BP (SAHRBF-BP) classifier.
  • To introduce an optimized learning strategy for adaptive network structure determination.
  • To enhance classification performance, particularly for low-dimensional, large-scale datasets.

Main Methods:

Related Experiment Videos

  • Developed a structure-adaptive RBF network that dynamically adjusts hidden nodes based on sample space distribution.
  • Integrated an optimized learning strategy using potential functions and repulsive forces to determine RBF hidden node parameters.
  • Cascaded the adaptive RBF network with a BP network, where RBF handles nonlinear mapping and BP performs nonlinear classification.

    • Main Results:

    • The SAHRBF-BP classifier adaptively determines the number of RBF hidden nodes and their parameters.
    • The number of BP input nodes is determined by the number of adaptively generated RBF hidden nodes.
    • Experimental results show the superiority of SAHRBF-BP over other algorithms on various datasets, especially low-dimensional, large ones.

    Conclusions:

    • The proposed SAHRBF-BP classifier offers an effective approach for adaptive nonlinear mapping and classification.
    • The optimized learning strategy significantly improves the determination of network structure and parameters.
    • SAHRBF-BP demonstrates enhanced classification performance compared to other SLFNs algorithms, particularly on challenging datasets.