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 Video

Updated: Jul 3, 2026

Augmenting Large Language Models via Vector Embeddings to Improve Domain-Specific Responsiveness
03:14

Augmenting Large Language Models via Vector Embeddings to Improve Domain-Specific Responsiveness

Published on: December 6, 2024

Enhancing clinically cardiovascular machine learning model for risk prediction via sample augmentation.

Xiaoyu Tang1,2,3, Min Tang4, Wu Liu5

  • 1Shenzhen Hospital (Fu Tian) of Guangzhou University of Chinese Medicine, The Sixth Clinical School, Guangzhou University of Chinese Medicine, Shenzhen, Guangdong, China.

Frontiers in Medicine
|June 25, 2026
PubMed
Summary

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

Prevalence of Dysphagia in China: An Epidemiological Survey of 5943 Participants.

Dysphagia·2020
Same author

Docetaxel suppresses immunotherapy efficacy of natural killer cells toward castration-resistant prostate cancer cells via altering androgen receptor-lectin-like transcript 1 signals.

The Prostate·2020
Same author

Drp1-dependent remodeling of mitochondrial morphology triggered by EBV-LMP1 increases cisplatin resistance.

Signal transduction and targeted therapy·2020
Same author

Solitary fibrous tumor involving urinary bladder: a case report and literature review.

Translational andrology and urology·2020
Same author

Sputum Cell-Free DNA: Valued Surrogate Sample for Detection of EGFR Mutation in Patients with Advanced Lung Adenocarcinoma.

The Journal of molecular diagnostics : JMD·2020
Same author

CD41-deficient exosomes from non-traumatic femoral head necrosis tissues impair osteogenic differentiation and migration of mesenchymal stem cells.

Cell death & disease·2020
This summary is machine-generated.

Moderate data augmentation, specifically 2x, improves machine learning models for cardiovascular risk prediction in small datasets. Random Forest (RF) offers the best balance of accuracy and interpretability for clinical deployment.

Area of Science:

  • Cardiovascular disease research
  • Machine learning in healthcare
  • Clinical data analysis

Background:

  • Small sample sizes and data heterogeneity challenge machine learning model robustness in clinical settings.
  • Developing reliable machine learning models for cardiovascular risk is crucial for patient outcomes.

Purpose of the Study:

  • To assess the impact of moderate data augmentation on cardiovascular risk modeling.
  • To propose an interpretable and deployable machine learning solution using a continuous risk to thresholding framework.

Main Methods:

  • Compared five machine learning models (SVR, RF, XGBoost, LightGBM, MLP) using a heart disease dataset split 8:2 for training/validation.
  • Applied constrained feature space augmentation (0x, 1x, 2x, 3x) and evaluated continuous risk scores using MAE, RMSE, and R².
Keywords:
SHAPcardiovascular riskmachine learningrandom forestrisk predictionsample augmentation

Related Experiment Videos

Last Updated: Jul 3, 2026

Augmenting Large Language Models via Vector Embeddings to Improve Domain-Specific Responsiveness
03:14

Augmenting Large Language Models via Vector Embeddings to Improve Domain-Specific Responsiveness

Published on: December 6, 2024

  • Utilized SHapley Additive exPlanations (SHAP) and partial dependence plots (PDP) for interpretability analysis.

    • Main Results:

    • 2x data augmentation optimized the balance between prediction error (reduced MAE/RMSE) and model fit (increased R²).
    • Random Forest (RF) demonstrated superior performance with 2x augmentation, achieving 94.0% accuracy, 94.4% F2-score, 95.9% sensitivity, and 91.8% specificity post-thresholding.
    • SHAP/PDP identified key risk factors (oldpeak, num major vessels, chest pain type, thal, exang, max hr) with stable predictive patterns.

    Conclusions:

    • Moderate data augmentation (2x) significantly enhances the robustness of machine learning models for cardiovascular risk in small datasets.
    • Random Forest (RF) provides an optimal combination of accuracy, stability, and interpretability.
    • The study offers a methodological framework for deploying interpretable cardiovascular risk models through multi-layered interpretation and thresholding.