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

Adaptive hybrid learning for neural networks.

Rob Smithies1, Said Salhi, Nat Queen

  • 1School of Mathematics and Statistics, University of Birmingham, UK. smithier@for.mat.bham.ac.uk

Neural Computation
|March 10, 2004
PubMed
Summary

This study introduces an enhanced Resilient Propagation (RPROP) algorithm, combining it with Local Search for improved machine learning classification. The hybrid method demonstrates superior speed and accuracy on natural datasets.

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

Development of an Optimized Process for Functional Recombinant SARS-CoV-2 Spike S1 Receptor-Binding Domain Protein Produced in the Baculovirus Expression Vector System.

Tropical medicine and infectious disease·2023
Same author

Immunohistochemical Characterization Improves the Reproducibility of the Histological Diagnosis of Ovarian Carcinoma

Asian Pacific journal of cancer prevention : APJCP·2018
Same author

Measuring the environmental sustainability performance of global supply chains: A multi-regional input-output analysis for carbon, sulphur oxide and water footprints.

Journal of environmental management·2016
Same author

Strategies for increasing the efficiency of a genetic algorithm for the structural optimization of nanoalloy clusters.

Journal of computational chemistry·2005

Area of Science:

  • Machine Learning
  • Optimization Algorithms

Background:

  • Gradient-based learning algorithms like Resilient Propagation (RPROP) have limitations.
  • Addressing these drawbacks is crucial for improving classification task performance.

Purpose of the Study:

  • To develop a robust, locally adaptive learning algorithm by enhancing the RPROP method.
  • To improve RPROP's performance by addressing its remaining drawbacks through hybridization.

Main Methods:

  • Two enhancements were applied to the standard Resilient Propagation (RPROP) algorithm.
  • Hybridization with gradient-independent Local Search was employed.
  • A global optimization method using recursion of the hybrid approach with tabu neighborhoods was constructed.

Main Results:

Related Experiment Videos

  • The enhanced RPROP algorithm demonstrated increased speed and accuracy compared to standard RPROP.
  • Performance was evaluated on classification tasks using natural datasets from the UCI repository.
  • The integration of Local Search within the global optimization method further improved Enhanced RPROP's effectiveness.

Conclusions:

  • The developed Enhanced RPROP algorithm offers a more robust and efficient solution for classification tasks.
  • Hybridization with Local Search and global optimization techniques significantly boosts learning algorithm performance.
  • The study validates the effectiveness of the proposed methods on real-world machine learning problems.