Hybrid Gradient Boosting and Graph Neural Network Architecture for Cardiovascular Disease Prediction

Kummetha Snehalatha, Mudhavarthi Phabe Zoe Gospel, Y. Indira Priyadarshini, Pasam Hima Teja, Kethapally Sri Sathwika

2025

Abstract

Heart disease (CVD) is one of the main causes of death around the world, so an advanced future model is required to assess sufficient risk. Although traditional machine learning techniques such as Gradient's Boosting Machines (GBMS) are effective with structured spot data, they cannot describe complex graph -based conditions usually found in health data. Conversely, Graph Neural Networks (GNNs) are suitable for graph-structured data but not tabular data. To fill this gap, this research proposes a hybrid model that combines GBM and GNN for CVD prediction, leveraging the strengths of both methods. Using the Framingham Heart Study dataset, a synthetic graph is built to capture patient similarities. Experimental results indicate that the proposed hybrid method outperforms individual models in accuracy and AUC-ROC, proving its ability to improve predictive performance and inform healthcare analytics using multimodal data fusion.

Paper Citation

in Harvard Style

Snehalatha K., Gospel M., Priyadarshini Y., Teja P. and Sathwika K. (2025). Hybrid Gradient Boosting and Graph Neural Network Architecture for Cardiovascular Disease Prediction. 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 287-291. DOI: 10.5220/0013881600004919

in Bibtex Style

@conference{icrdicct`2525,
author={Kummetha Snehalatha and Mudhavarthi Gospel and Y. Priyadarshini and Pasam Teja and Kethapally Sathwika},
title={Hybrid Gradient Boosting and Graph Neural Network Architecture for Cardiovascular Disease Prediction},
booktitle={Proceedings of the 1st International Conference on Research and Development in Information, Communication, and Computing Technologies - ICRDICCT`25},
year={2025},
pages={287-291},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0013881600004919},
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 Gradient Boosting and Graph Neural Network Architecture for Cardiovascular Disease Prediction
SN - 978-989-758-777-1
AU - Snehalatha K.
AU - Gospel M.
AU - Priyadarshini Y.
AU - Teja P.
AU - Sathwika K.
PY - 2025
SP - 287
EP - 291
DO - 10.5220/0013881600004919
PB - SciTePress