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

The recursive deterministic perceptron neural network.

Mohamed Tajine1, David Elizondo

  • 1LSIIT (CNRS URA 1871), Université Louis Pasteur, Département d'Informatique, 7 rue René Descartes, 67084, Strasbourg, France

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

Urinary Metabolic Distinction of Niemann-Pick Class 1 Disease through the Use of Subgroup Discovery.

Metabolites·2023
Same author

Application of data augmentation techniques towards metabolomics.

Computers in biology and medicine·2022
Same author

Dealing with distribution mismatch in semi-supervised deep learning for COVID-19 detection using chest X-ray images: A novel approach using feature densities.

Applied soft computing·2022
Same author

A real use case of semi-supervised learning for mammogram classification in a local clinic of Costa Rica.

Medical & biological engineering & computing·2022
Same author

Correcting data imbalance for semi-supervised COVID-19 detection using X-ray chest images.

Applied soft computing·2021
Same author

Design of robust vascular tree matching: validation on liver.

Information processing in medical imaging : proceedings of the ... conference·2007
Same journal

A boundary-regularization-enhanced video anomaly detection network based on context-adaptive spatio-temporal conditional diffusion.

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

MT<sup>2</sup>-CSD and LLM-CRAN: A new dataset and an LLM-based multi-semantic knowledge fusion model for conversational stance detection.

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

TriAlignNet: A triple-path cross-modality alignment framework for multimodal time series forecasting.

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

Anchor-based disentanglement framework for incremental multi-view clustering.

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

Complex-valued amplitude-phase interference modeling for adversarially robust classification.

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

TraNce: Type-aware hypergraph neural network with biological mediators for drug repositioning.

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

We introduce the recursive deterministic perceptron (RDP), a neural network model that solves any two-class classification problem, unlike the single-layer perceptron. Growing methods automatically construct the RDP, enabling it to handle complex, non-linearly separable data.

Area of Science:

  • Machine Learning
  • Artificial Intelligence
  • Computational Neuroscience

Background:

  • Single-layer perceptrons (SLPT) are limited to linearly separable classification problems.
  • Complex datasets often require more sophisticated classification models.
  • Existing growing methods for neural networks have varying performance characteristics.

Purpose of the Study:

  • Introduce a novel feedforward multilayer neural network, the recursive deterministic perceptron (RDP).
  • Develop RDPs capable of solving any two-class classification problem, overcoming SLPT limitations.
  • Generalize the RDP for multi-class (m>2) classification and function approximation.

Main Methods:

  • Proposed several growing methods to construct RDPs by adding intermediate neurons (INs).

Related Experiment Videos

  • Each IN augments input vectors with its output, increasing affine dimension.
  • Demonstrated RDP construction by iteratively selecting linearly separable (LS) subsets.
  • Introduced a generalized RDP for m-class classification based on a new notion of linear separability.

    • Main Results:

    • RDPs can solve any two-class classification problem, including non-linearly separable ones.
    • Growing methods automatically determine RDP topology and weights.
    • The construction of a new IN is proven NP-complete under maximum cardinality assumptions.
    • The generalized RDP effectively handles m-class classification and function approximation.
    • Comparative analysis showed RDP performance against backpropagation (BP) and cascade correlation (CC).

    Conclusions:

    • The RDP model offers a powerful alternative to traditional perceptrons for complex classification tasks.
    • Automated growing methods provide an efficient way to construct effective RDP topologies.
    • The RDP generalization extends its utility to multi-class problems and continuous function approximation.
    • RDPs demonstrate competitive or superior performance compared to established neural network models in classification benchmarks.