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

On global-local artificial neural networks for function approximation.

David Wedge, David Ingram, David McLean

    IEEE Transactions on Neural Networks
    |July 22, 2006
    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

    Electroencephalography, behavioural observations, bleed-out and temperature at the point of application in cattle receiving DTS: Diathermic Syncope® at 160-200 kJ.

    Veterinary and animal science·2026
    Same author

    Red Blood Cell Exchange Transfusion for Severe Babesiosis.

    JAMA internal medicine·2026
    Same author

    Single base focal hypermutation cooccurs with structural variation as an early event in advanced prostate tumourigenesis with ancestry specific independence: a multi-ancestral observational study.

    Research square·2025
    Same author

    Screening of Babesia microti Proteomic Microarrays With Acute and Convalescent Serum of Patients Reveals Novel Targets of Early IgM and IgG Antibodies.

    The Journal of infectious diseases·2025
    Same author

    Cattle recover completely from unconsciousness induced by controlled application of 150-180 kJ of 915 MHz microwave energy to the forehead.

    Veterinary and animal science·2025
    Same author

    Tissue integrity impacts of application of 160-200 kJ of 915 MHz microwave energy, using the DTS: Diathermic Syncope® system, to the forehead of cattle, and alignment with the requirements of religious slaughter markets.

    Veterinary and animal science·2025
    Same journal

    Universal perceptron and DNA-like learning algorithm for binary neural networks: LSBF and PBF implementations.

    IEEE transactions on neural networks·2013
    Same journal

    Guest editorial: special section on white box nonlinear prediction models.

    IEEE transactions on neural networks·2011
    Same journal

    Data-based fault-tolerant control of high-speed trains with traction/braking notch nonlinearities and actuator failures.

    IEEE transactions on neural networks·2011
    Same journal

    Guest editorial: special section on data-based control, modeling, and optimization.

    IEEE transactions on neural networks·2011
    Same journal

    Neural network-based multiple robot simultaneous localization and mapping.

    IEEE transactions on neural networks·2011
    Same journal

    Data-driven model-free adaptive control for a class of MIMO nonlinear discrete-time systems.

    IEEE transactions on neural networks·2011
    See all related articles

    A new hybrid radial basis function (RBF) sigmoid neural network, the global-local artificial neural network (GL-ANN), efficiently approximates functions by separating global and local features. This approach achieves lower errors and requires fewer neurons than single-type networks.

    Area of Science:

    • Artificial Intelligence
    • Machine Learning
    • Computational Science

    Background:

    • Function approximation is crucial in various scientific and engineering domains.
    • Traditional neural network architectures often struggle with efficiently capturing both global and local patterns in data.
    • Radial basis function (RBF) networks and sigmoid neural networks are established methods, but hybrid approaches offer potential improvements.

    Purpose of the Study:

    • To introduce a novel hybrid radial basis function (RBF) sigmoid neural network architecture.
    • To develop a three-step training algorithm that combines global search and gradient descent for efficient function approximation.
    • To demonstrate the effectiveness of the proposed global-local artificial neural network (GL-ANN) in handling complex input-output relationships.

    Main Methods:

    Related Experiment Videos

    • A hybrid neural network combining RBF and sigmoid neurons is proposed.
    • A three-step training algorithm is employed, prioritizing global feature identification before local detail refinement.
    • The method was evaluated on five regression tasks, including synthetic datasets and real-world wave overtopping data.

    Main Results:

    • The hybrid GL-ANN architecture demonstrated superiority over single-type neuron networks in several regression tasks.
    • Achieved lower mean square errors with fewer hidden neurons and reduced need for regularization.
    • Outperformed perceptron RBF nets and regression tree-derived RBFs in comparative analyses.

    Conclusions:

    • The proposed GL-ANN offers an efficient and effective approach to function approximation.
    • The hybrid architecture successfully balances global and local feature identification for improved performance.
    • Further research is warranted on training aspects like regularization, optimization, and model selection for GL-ANNs.