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

Improving diagnostic accuracy using a hierarchical neural network to model decision subtasks.

D West1, V West

  • 1Department of Decision Sciences, College of Business Administration, East Carolina University, Greenville, NC 27836, USA. westd@mail.ecu.edu

International Journal of Medical Informatics
|March 9, 2000
PubMed
Summary

This study enhanced diagnostic accuracy for erythematous-squamous diseases using hierarchical neural networks. A two-stage model achieved performance close to the theoretical best, significantly improving disease diagnosis.

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

Revisiting the Formulation of Charged Defect in Solids.

Physical review letters·2025
Same author

Admission screening testing of patients and staff N95 respirators are cost-effective in reducing COVID-19 hospital-acquired infections.

The Journal of hospital infection·2024
Same author

Current status of cystic echinococcosis control in the Falkland Islands: has elimination been achieved?

Parasitology·2023
Same author

Safety of short 3-hour recovery and same-day discharge following lower-limb angioplasty in outpatients with intermittent claudication and critical limb ischaemia.

Clinical radiology·2022
Same author

The National Lung Cancer Audit: The Impact of COVID-19.

Clinical oncology (Royal College of Radiologists (Great Britain))·2022
Same author

iPSC modeling of young-onset Parkinson's disease reveals a molecular signature of disease and novel therapeutic candidates.

Nature medicine·2020

Area of Science:

  • Medical Informatics
  • Artificial Intelligence in Medicine
  • Dermatology

Background:

  • Quantitative models like neural networks are used in medical diagnostic support systems.
  • Accurate differential diagnosis of erythematous-squamous diseases remains a challenge.

Purpose of the Study:

  • To investigate the diagnostic accuracy of neural network models for six erythematous-squamous diseases.
  • To explore hierarchical neural network models for improving diagnostic accuracy by partitioning decision domains.
  • To design and evaluate a two-stage hierarchical neural network for differential diagnosis.

Main Methods:

  • Utilized various quantitative models including neural networks for disease diagnosis.
  • Employed Self-Organizing Maps (SOM) to visualize 34 feature variables and identify inconsistent cases.

Related Experiment Videos

  • Developed a two-stage hierarchical neural network combining a multilayer perceptron and a mixture-of-experts model.

    • Main Results:

    • SOM analysis identified five inconsistent cases, establishing a lower bound for diagnostic error at 0.0140.
    • Traditional quantitative models resulted in error levels substantially greater than the target.
    • The designed two-stage hierarchical neural network achieved a diagnostic error rate of 0.0159, approaching the SOM-established target.

    Conclusions:

    • Hierarchical neural network models can enhance diagnostic accuracy by breaking down complex decision domains.
    • The developed two-stage hierarchical neural network shows significant promise for the differential diagnosis of erythematous-squamous diseases.
    • This approach offers a substantial improvement over traditional quantitative diagnostic models.