Ensemble-Based Intrusion Detection for IoT Networks Using the CICIoT2023 Dataset

Abstract

The rapid proliferation of Internet of Things (IoT) devices has enhanced automation and connectivity across industries, including healthcare, smart homes, and industrial systems. However, the interconnected nature of IoT networks also exposes them to significant cybersecurity threats, making them a prime target for cyberattacks. Intrusion Detection Systems (IDS) play a crucial role in securing these networks by identifying and mitigating potential threats. This research explores machine learning-based intrusion detection techniques tailored for IoT networks, utilizing models such as Decision Trees, Random Forests, Support Vector Machines (SVM), and Neural Networks to classify network traffic as benign or malicious. A comprehensive dataset with key features, including protocol types, packet flags, and flow statistics, was used for model training and evaluation. The study focuses on enhancing threat detection capabilities while maintaining a balance between performance and computational efficiency. The proposed approach demonstrates the effectiveness of machine learning-driven IDS in strengthening IoT security by reducing false positives and ensuring reliable detection of malicious activity. Additionally, the research highlights challenges in real-time intrusion detection, the importance of feature selection, and strategies for optimizing IDS for resource-constrained IoT environments. The findings contribute to the development of adaptive and scalable intrusion detection solutions, paving the way for more resilient IoT ecosystems.