Detection and Classification of Dos Attacks in Zigbee Networks Using Supervised Learning Algorithms

Abstract

[Abstract] The Zigbee protocol, designed for low-power wireless personal area networks, is a key technology for the Internet of Things (IoT). This paper explores detecting denial-of-service (DoS) attacks in Zigbee networks through supervised classification methods. A specialized dataset was generated from a simulated Zigbee environment, capturing both normal and malicious traffic. The study evaluates the effectiveness of six supervised classification algorithms, including Logistic Regression, K-Nearest Neighbors (KNN), Support Vector Machines (SVM), Decision Trees, Random Forests, and Artificial Neural Networks (ANN). Results indicate that KNN, Decision Trees, and Random Forests achieved superior performance with high accuracy and efficiency. This methodology highlights the potential of tailored datasets and machine learning approaches for enhancing security in Zigbee networks and protecting IoT environments from evolving threats.