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

Comparing error minimized extreme learning machines and support vector sequential feed-forward neural networks.

Enrique Romero1, René Alquézar

  • 1Departament de Llenguatges i Sistemes Informàtics, Universitat Politècnica de Catalunya, Barcelona, Spain. eromero@lsi.upc.edu

Neural Networks : the Official Journal of the International Neural Network Society
|October 1, 2011
PubMed
Summary
This summary is machine-generated.

Related Concept Videos

Mechanistic Models: Compartment Models in Algorithms for Numerical Problem Solving01:29

Mechanistic Models: Compartment Models in Algorithms for Numerical Problem Solving

Mechanistic models play a crucial role in algorithms for numerical problem-solving, particularly in nonlinear mixed effects modeling (NMEM). These models aim to minimize specific objective functions by evaluating various parameter estimates, leading to the development of systematic algorithms. In some cases, linearization techniques approximate the model using linear equations.
In individual population analyses, different algorithms are employed, such as Cauchy's method, which uses a...

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by
Same author

Behavioral and structural drivers of HIV in men who have sex with men in Belize: a mixed-methods study.

Revista panamericana de salud publica = Pan American journal of public health·2026
Same author

The fascinating complexity of seagrass bio-fibres: insights from bio-chemo-hygro mechanical analysis for their reuse as soil reinforcement.

Scientific reports·2026
Same author

Machine Learning Prediction of Cardiovascular Risk in Type 1 Diabetes Mellitus Using Radiomic Features from Multimodal Retinal Images.

Ophthalmology science·2025
Same author

Mean Field Initialization of the Annealed Importance Sampling Algorithm for an Efficient Evaluation of the Partition Function Using Restricted Boltzmann Machines.

Entropy (Basel, Switzerland)·2025
Same author

Radiomics-Based Assessment of OCT Angiography Images for Diabetic Retinopathy Diagnosis.

Ophthalmology science·2022
Same author

On-The-Fly Syntheziser Programming with Fuzzy Rule Learning.

Entropy (Basel, Switzerland)·2020

Support vector sequential feed-forward neural networks (SV-SFNNs) show improved generalization performance over error minimized extreme learning machines (EM-ELMs). Both methods construct single-hidden-layer feed-forward networks sequentially with similar computational costs.

Area of Science:

  • Computational Intelligence
  • Machine Learning
  • Neural Networks

Background:

  • Error Minimized Extreme Learning Machines (EM-ELMs) offer an efficient sequential approach for building single-hidden-layer feed-forward networks (SLFNs).
  • Existing sequential SLFN construction methods include Support Vector Sequential Feed-Forward Neural Networks (SV-SFNNs), a specific case of Sequential Approximation with Optimal Coefficients and Interacting Frequencies (SAOCIF), which utilize data subsets for hidden layers instead of random weights.

Purpose of the Study:

  • To demonstrate that EM-ELMs can be viewed as a specific instance of the SAOCIF method.
  • To enhance EM-ELMs by incorporating SAOCIF's computational strategies for output-layer weights and candidate selection.
  • To experimentally compare the generalization performance of EM-ELMs and SV-SFNNs under identical conditions.

Related Experiment Videos

Main Methods:

  • Theoretical analysis to establish EM-ELMs as a special case of SAOCIF.
  • Modification of EM-ELMs to integrate SAOCIF's methods for selecting random hidden node candidates and optimizing output weights.
  • Empirical evaluation using 10 benchmark classification and 10 benchmark regression datasets.

Main Results:

  • EM-ELMs were successfully reformulated as a particular case of SAOCIF, allowing for potential improvements in candidate testing and output weight computation.
  • Experimental results indicated that SV-SFNNs achieved statistically significant improvements in generalization performance compared to EM-ELMs.
  • This performance advantage for SV-SFNNs was observed in 12 out of the 20 tested benchmark datasets, despite both models having comparable computational efficiency.

Conclusions:

  • EM-ELMs share theoretical underpinnings with the SAOCIF framework, suggesting avenues for algorithmic enhancement.
  • SV-SFNNs demonstrate superior generalization capabilities compared to standard EM-ELMs on a variety of benchmark tasks.
  • The findings highlight the benefits of using data-driven hidden layer weights (as in SV-SFNNs) over random weights (as in EM-ELMs) for sequential SLFN construction.