Enhancing Security and Privacy in Blockchain-IoT Hybrid Healthcare Database Systems through Federated Learning
S. Loganatha Prasanna, K. Himabindu, P. Sowjanya, Yuvraj Dilip Patil, Mukesh Soni, Anil Kumar Lamba · 2024
This study focusses on the crucial matter of security and privacy in healthcare database systems, specifically in the context of blockchain-IoT hybrid ecosystems. Conventional machine learning methods, although efficient, sometimes fail to adequately protect sensitive healthcare data, particularly in the context of decentralized systems. A new solution is suggested to address this issue by combining federated learning with blockchain-IoT technologies. This integration provides improved data security and privacy while maintaining high speed. The suggested methodology utilizes decentralized learning to guarantee that data remains localized, hence minimizing the possibility of breaches and unauthorized access. The suggested system outperforms existing approaches such as Random Forest and Support Vector Machine (SVM), reaching an accuracy of 97.8%, an F1 score of 95.85%, and a recall of 95.2%. The model's performance in safe and accurate data processing is demonstrated by its findings, which exceed the metrics of Random Forest (93.2% accuracy) and SVM (95.4% accuracy). The research not only enhances the sector by delivering a more resilient solution but also paves the way for practical applications in healthcare, presenting a substantial enhancement over current technology.