Federated learning for intrusion detection in IoT security: a hybrid ensemble approach

Critical role of the internet things (IoT) in various domains like smart city, healthcare, supply chain, and transportation has made them target malicious attacks. Past works this area focused on centralised intrusion detection system (IDS), assuming a central entity to perform data analysis identify threats. However, such IDS may not always be feasible, mainly due spread across multiple sources, gathering at node can costly. In paper, we first present an architecture for based hybrid ensemble model named PHEC, which gives improved performance compared state-of-the-art architectures. We then adapt federated learning framework. Next, propose noise-tolerant PHEC address label-noise problem. Experimental results four benchmark datasets drawn from security attacks show that our achieves high TPR while keeping FPR low noisy clean data.