Heart Disease Prediction Using Federated Learning

A pioneering approach to predicting heart disease through the implementation of federated learning (FL), a collaborative machine learning framework is addressed. FL works by training a central model across decentralized devices or servers while preserving the privacy of healthcare dataset (s). FL involves the cooperation of various clients, spanning individual devices to entire organizations, under the centralized orchestration of a server. The model aims to address challenges related to privacy and communication efficiency in the healthcare sector, specifically focusing on predicting cardiovascular risk factors. The paper introduces a robust FL environment and utilizes a diverse dataset sourced from the UCI Machine Learning Repository, ensuring exposure to a wide range of relevant features and scenarios. The model’s adaptability is demonstrated by training it on both a local system and a local area network (LAN). Real-world deployment scenarios showcase the model’s effectiveness, achieving an impressive accuracy rate of up to 89.98% in predicting cardiovascular risk.