Strengthening IoT Security: Leveraging Machine Learning for Improved Detection of Intrusions in Connected Networks

EasyChair Preprint 12486

Abstract

The rapid proliferation of Internet of Things (IoT) devices has led to unprecedented connectivity, revolutionizing various aspects of our lives. However, this interconnectedness also brings forth significant security challenges, as IoT devices often lack robust built-in security measures. In this context, effective intrusion detection systems (IDS) are crucial for safeguarding IoT networks against malicious attacks. This paper proposes a novel approach to fortifying IoT security by harnessing the power of machine learning for enhanced intrusion detection. By leveraging machine learning algorithms, such as anomaly detection and supervised classification, our system aims to accurately identify and mitigate potential intrusions in interconnected IoT networks. Unlike traditional rule-based IDS, which may struggle to adapt to evolving threats and complex network behaviors, our approach offers the flexibility to dynamically learn and adapt to new attack patterns. Through comprehensive experimentation and evaluation on real-world IoT datasets, we demonstrate the effectiveness and scalability of our proposed system in detecting various types of intrusions while minimizing false positives. By integrating machine learning techniques into IoT security frameworks, we strive to provide a proactive and robust defense mechanism against emerging cyber threats, thus fostering a safer and more secure IoT ecosystem for users and stakeholders alike.

Keyphrases: Cybersecurity, Intrusion Detection, IoT Security, Network Defense, Threat Detection, anomaly detection, interconnected networks, machine learning, supervised classification

@booklet{EasyChair:12486,
  author    = {Battle Hurry},
  title     = {Strengthening IoT Security: Leveraging Machine Learning for Improved Detection of Intrusions in Connected Networks},
  howpublished = {EasyChair Preprint 12486},
  year      = {EasyChair, 2024}}