An Explainable Intrusion Detection System for IoT Networks
Michela Fazzolari, Pietro Ducange, Francesco Marcelloni
Abstract
The security of Internet of Things (IoT) networks is a pressing concern, as these networks are vulnerable to malicious attacks that can result in serious consequences. In this paper, we present a novel explainable Intrusion Detection System (IDS) capable of discriminating authentic from malicious network traffic within a IoT network of smart devices. The system adopts a Fuzzy Decision Tree as an eXplainable Artificial Intelligence (XAI) model for actually classifying the IoT network traffic. We evaluate the effectiveness of our approach considering the simulated attacks carried out by 3 devices of an IoT network, previously infected by a botnet. Preliminary results show that the proposed IDS, based on fuzzy decision trees, achieves promising results in terms of both explainability and ability to distinguish authentic traffic from 5 different types of malicious network traffic.