Federated Learning: A Comparative Survey on Privacy-Preserving Approach to Medical Intelligence Models

Abstract

Federated Learning (FL) is a transformative approach to machine learning that enables collaborative model training across multiple entities without compromising data privacy. In healthcare, where patient data is highly sensitive and governed by stringent regulations such as The Health Insurance Portability and Accountability Act (HIPAA) in the United States and General Data Protection Regulation (GDPR) in the European Union, FL offers a privacy-preserving solution. This study investigates the application of FL in predicting cancer outcomes, comparing its performance against traditional machine learning algorithms, including Logistic Regression, based on key metrics such as accuracy, precision, recall, F1-score, and training time. The results demonstrate that FL outperforms Logistic Regression with an accuracy of 89%, precision of 88.67%, recall of 86%, and an F1-score of 89.7%, while maintaining competitive training efficiency. This paper also provides practical implementations of FL in real-world healthcare scenarios, showcasing its potential to address privacy challenges and enable robust medical data analysis. By leveraging FL, healthcare institutions can achieve enhanced collaboration, improved predictive accuracy, and compliance with data protection regulations, paving the way for advancements in privacy-sensitive medical machine learning applications.