Privacy-Preserving Heart Disease Prediction Using TabTransformer-Based Federated Learning on Non-IID Clinical Data

View at publisher Summarise with VIRAT
Authors: Chamudika Pallewela
Cited by: 0
References: 13

Summary

No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.

Summarise with VIRAT Free to try with your Science Catalogs account.

Abstract

Heart disease remains a leading cause of mortality worldwide, and accurate early prediction is critical for timely clinical intervention. Traditional machine learning and deep learning approaches often rely on centralised patient data, raising significant concerns regarding privacy, security, and regulatory compliance. In this study, the researcher proposes a privacy-preserving federated learning framework that integrates a TabTransformer architecture with the FedProx optimisation algorithm to predict heart disease from tabular clinical data. The framework is specifically designed to address non-identically distributed (non-IID) client data across multiple healthcare institutions, ensuring stable model convergence while preserving data confidentiality. Comprehensive evaluations including accuracy, F1-score, precision, recall, ROC-AUC, and loss—demonstrate that the proposed approach outperforms conventional federated models and achieves robust performance at both client and global levels. These findings underscore the potential of transformer-based federated learning as a scalable, privacy-preserving solution for heart disease prediction, paving the way for real-world deployment in clinical environments.In addition to demonstrating strong predictive accuracy, the proposed framework is benchmarked against baseline FL and centralised models, showing improved generalisation under heterogeneous conditions. A privacy-preserving protocol and interpretability analysis further highlight its practicality for real-world deployments.

Fields of Study & Subjects

Privacy-Preserving Technologies in Data Artificial Intelligence in Healthcare Machine Learning in Healthcare

Publication details

Journal / source
IEEE International Conference on Industrial and Information Systems (ICIIS)
Pages
454–459
Publication date
16 January 2026
Language
EN