Hybrid Machine Learning Model for Early Prediction of Coronary Artery Disease: Integrating LightGBM and Ensemble Techniques for Enhanced Accuracy

Kavin Kumar D., Devendhiran S., Gomathy G., Karnish N.

2025

Abstract

Coronary Artery Disease (CAD) is a major issue confronting the global community today, which cannot be foregone without some accurate predictive models for early diagnosis and intervention. In this study, therefore, a Hybrid Machine Learning Model, HY OptGBM-Ensemble, is designed to combine LightGBM with techniques for ensemble aiming at higher accuracy in predictions while dealing with class imbalance. The model uses Optima-based hyperparameter tuning along with focal loss optimization and recursive feature elimination to fine- tune feature selection and improve classification. Comparative evaluation against LightGBM, XGBoost, and Logistic Regression shows that the proposed hybrid model has obtained AUC scores 97.8% on the Framingham dataset. Class distinction is also improved along with predictive capability. SHAP analysis further increases model interpretability.

Paper Citation

in Harvard Style

D. K., S. D., G. G. and N. K. (2025). Hybrid Machine Learning Model for Early Prediction of Coronary Artery Disease: Integrating LightGBM and Ensemble Techniques for Enhanced Accuracy. In Proceedings of the 1st International Conference on Research and Development in Information, Communication, and Computing Technologies - ICRDICCT`25; ISBN 978-989-758-777-1, SciTePress, pages 525-530. DOI: 10.5220/0013932400004919

in Bibtex Style

@conference{icrdicct`2525,
author={Kavin D. and Devendhiran S. and Gomathy G. and Karnish N.},
title={Hybrid Machine Learning Model for Early Prediction of Coronary Artery Disease: Integrating LightGBM and Ensemble Techniques for Enhanced Accuracy},
booktitle={Proceedings of the 1st International Conference on Research and Development in Information, Communication, and Computing Technologies - ICRDICCT`25},
year={2025},
pages={525-530},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0013932400004919},
isbn={978-989-758-777-1},
}

in EndNote Style

TY - CONF

JO - Proceedings of the 1st International Conference on Research and Development in Information, Communication, and Computing Technologies - ICRDICCT`25
TI - Hybrid Machine Learning Model for Early Prediction of Coronary Artery Disease: Integrating LightGBM and Ensemble Techniques for Enhanced Accuracy
SN - 978-989-758-777-1
AU - D. K.
AU - S. D.
AU - G. G.
AU - N. K.
PY - 2025
SP - 525
EP - 530
DO - 10.5220/0013932400004919
PB - SciTePress