Attack Detection in IoT-Based Healthcare Networks Using Hybrid Federated Learning

Cybercrimes are increasing rapidly throughout the world, leading to financial losses and compromising the integrity and confidentiality of private data. Statistics showed that cybercrimes led to losses of around $6 trillion in 2021 based on a survey by Cybersecurity Ventures. Knowing that IoT networks are considered a source of identifiable data for vicious attackers to carry out criminal actions using automated processes, machine learning (ML)-assisted methods for IoT security have gained much attention in recent years. While conventional ML relies on a single server to store all of its data, which makes it a less desirable option for domains concerned about user privacy, the Federated Learning (FL)-based anomaly detection technique, which utilizes decentralized on-device data to identify IoT network intrusions, represents the proposed solution to the aforementioned problem. We propose a framework to train and test IoT data from health network using different classical machine learning algorithms and an enhanced federated learning model. FL is a framework that learns continuously in an iterative manner by training locally at the client side with the clientś individual data, and then updating the central server by forwarding the required data. We evaluated the performance of different algorithms based on accuracy, precision, recall and F1-score via different iterations. To develop a strong detection system, we used multiple datasets and generated different results. These results show decent and promising accuracy hence a promising solution towards telehealth application using machine learning techniques in detecting threats on IoT networks.