An Intrusion Detection System for IoT Using KNN and Decision-Tree Based Classification

Abstract

The Internet of Things (IoT) has grown rapidly in recent years, intending to affect everything from everyday life to large industrial systems. Regrettably, this has attracted the attention of hackers, who have turned the Internet of Things into a target for malicious activity, exposing end nodes to attack. IoT devices’ sheer volume and diversity make protecting the IoT infrastructure with a traditional intrusion detection system difficult. So to protect IoT devices, the data flow was investigated in an IoT context to protect these devices from hackers. We used two machine learning classifiers in this work: KNN (K-Nearest Neighbors) and DT (Decision Tree). We calculated the Error Rate, Accuracy, Precision, Recall, and F1 score for each method. When we combined these two classifiers, we obtained outstanding results (100 %). We have a high rate of detection of attacks. The findings are summarized.