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 machine learning performance by removing redundant cases in medical data sets

L Ohno-Machado1, H S Fraser, A Ohrn

  • 1Decision Systems Group, Brigham and Women's Hospital, Harvard Medical School, Boston, MA, USA. machado@dsg.harvard.edu

Proceedings. AMIA Symposium
|February 3, 1999
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

FHIRing up OpenMRS: Architecture, Implementation and Real-World Use-Cases in Global Health.

AMIA Joint Summits on Translational Science proceedings. AMIA Joint Summits on Translational Science·2024
Same author

Training in Health Informatics in Brazil.

Yearbook of medical informatics·2016
Same author

Creating a Common Data Model for Comparative Effectiveness with the Observational Medical Outcomes Partnership.

Applied clinical informatics·2015
Same author

Health IT and clinical decision support systems: human factors and successful adoption.

Journal of the American Medical Informatics Association : JAMIA·2014
Same author

"Big data" and the electronic health record.

Yearbook of medical informatics·2014
Same author

Confluence of disciplines in health informatics: an international perspective.

Methods of information in medicine·2011

Removing up to 86% of redundant cases from neural network training datasets for myocardial infarction classification did not significantly impact performance. This approach substantially reduces training time for machine learning models in healthcare.

Area of Science:

  • Medical informatics
  • Machine learning in healthcare

Background:

  • Neural network models and machine learning are effective for medical classification.
  • Periodic retraining of models with new data is beneficial.
  • Training neural networks can be time-consuming, necessitating efficient training sets.

Purpose of the Study:

  • To investigate the impact of removing redundant cases from training datasets on neural network classification performance.
  • To determine if significant classification performance is affected by data reduction.
  • To assess the potential for reducing neural network training time through data optimization.

Main Methods:

  • Experiments were conducted on a dataset of 700 patients with suspected myocardial infarction.
  • Redundant cases were randomly removed from the training set, with up to 86% of data being removed.

Related Experiment Videos

  • Classification performance was measured using the area under the Receiver Operating Characteristic (ROC) curve on independent test sets.

    • Main Results:

    • No statistically significant difference in classification performance was observed even when up to 86% of cases were removed.
    • Performance was evaluated on two previously unseen datasets of 553 and 500 cases.
    • A proportional reduction in the time required to train the neural network model was achieved.

    Conclusions:

    • Redundant case removal is a viable strategy for optimizing neural network training datasets in medical classification.
    • Significant data reduction is possible without compromising diagnostic accuracy for myocardial infarction.
    • This method offers a practical approach to accelerate the development and deployment of machine learning models in clinical settings.