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

Stability problems with artificial neural networks and the ensemble solution.

P Cunningham1, J Carney, S Jacob

  • 1Department of Computer Science, Trinity College, Dublin, Ireland. padraig.cunningham@tcd.ie

Artificial Intelligence in Medicine
|September 22, 2000
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

Optimized cellular perfusate increases lung viability ex vivo and extends the duration of ex vivo lung perfusion.

The Journal of heart and lung transplantation : the official publication of the International Society for Heart Transplantation·2026
Same author

Genetic Whispers in Hyperphosphataemic: Tumor Calcinosis in a 30-year-old.

Irish medical journal·2026
Same author

Bioconversion of aflatoxin-contaminated groundnut press cake by larvae of black soldier fly Hermetia illucens results in a complete mass balance for aflatoxin B<sub>1</sub>.

NPJ science of food·2024
Same author

The emerging role of lipid nanosystems and nanomicelles in liver diseases.

European review for medical and pharmacological sciences·2023
Same author

Something to SMILE about. Is small incision lenticule extraction ready to become the gold standard in laser refractive surgery? Yes.

Eye (London, England)·2023
Same author

Something to SMILE about. Is small incision lenticule extraction (SMILE) ready to become the gold standard in laser refractive surgery? no.

Eye (London, England)·2023
Same journal

Real-time EEG-based epileptic seizure prediction using artificial intelligence: A systematic review.

Artificial intelligence in medicine·2026
Same journal

R-peak detection and ECG data compression scheme based on empirical mode decomposition and wavelet transform.

Artificial intelligence in medicine·2026
Same journal

CastNet: A three-channel EEG-based deep learning model for cross-subject depression detection.

Artificial intelligence in medicine·2026
Same journal

State-of-the-art TinyML approaches for colorectal cancer detection: Current advances, challenges, and future directions.

Artificial intelligence in medicine·2026
Same journal

JRadiEvo: A Japanese radiology report generation model enhanced by evolutionary optimization of model merging.

Artificial intelligence in medicine·2026
Same journal

Causally-informed deep learning towards explainable and generalizable outcome prediction in critical care.

Artificial intelligence in medicine·2026
See all related articles

Artificial neural networks (ANNs) are unstable predictors in medical systems. For in-vitro fertilisation, k-fold cross-validation and aggregating predictors improve accuracy and generalization.

Area of Science:

  • Medical Informatics
  • Machine Learning
  • Reproductive Medicine

Background:

  • Artificial neural networks (ANNs) are widely used in medical decision support systems for classification and regression.
  • A significant limitation of ANNs is their inherent instability, where minor variations in training data can lead to substantially different models.
  • This instability directly impacts the generalizability and accuracy of predictive models.

Purpose of the Study:

  • To demonstrate the instability of ANNs in a specific medical application: in-vitro fertilisation (IVF) prediction.
  • To advocate for rigorous validation methods for ANN-based medical predictors.
  • To present a strategy for enhancing the accuracy of ANN predictors through aggregation.

Main Methods:

  • Detailed analysis of ANN behavior with varying training datasets in an IVF prediction context.

Related Experiment Videos

  • Application of k-fold cross-validation as a robust testing methodology.
  • Ensemble methods involving the aggregation of outputs from multiple ANN predictors.

    • Main Results:

    • Empirical evidence showing how small changes in training data for ANNs can lead to significant differences in model performance for IVF prediction.
    • Demonstration that k-fold cross-validation provides a more reliable assessment of generalization accuracy compared to single-train/test splits.
    • Significant improvements in prediction accuracy achieved by combining the outputs of several aggregated ANN models.

    Conclusions:

    • Claims regarding the generalization performance of ANNs in medical applications, particularly IVF, must be substantiated by k-fold cross-validation.
    • Aggregating multiple ANN predictors is an effective technique to mitigate instability and enhance predictive accuracy.
    • The findings underscore the need for careful model validation and ensemble strategies in developing reliable AI-driven medical decision support systems.