Safeguarding IoT Networks Using Machine Learning for Intrusion Detection & Prevention
Open Access DepositedAs the Internet of Things (IoT) expands, more sophisticated security measures are needed to safeguard IoT networks due to increasing cyber threats. Statistical research highlights the IoT sector's rapid expansion, with forecasts of $1.4 trillion in investment by the end of 2023 and 21.5 billion devices globally in the same year. By 2030, 29.42 billion connected devices will be on the market, a 94.3% increase from 2023. IoT adoption has surged, but cybersecurity is still critical. IoT devices accounted for 66% of all devices in 2023, a sharp increase from 9% in 2010. Since 2018, investments in IoT have increased by 54.8%, reaching $1 trillion. The IoT market is predicted to reach $3.35 trillion by 2030; 2025, data creation is expected to reach 79.4 zettabytes. According to corporate trends, 34% of businesses have implemented IoT, and 94% of executives see advantages over risks. I.T. The ever-growing number of IoT devices has resulted in never-before-seen challenges regarding network security. To protect IoT networks from intrusions, it is critical to investigate Machine Learning (ML) models as proactive and adaptable solutions as the attack surface for cyber threats explodes, making existing security methods insufficient. The present research explores the security vulnerabilities in Intrusion Detection Systems (IDS) in IoT networks, focusing on how susceptible they are to cyberattack disruptions—using supervised learning and unsupervised ML-based approaches to detect intrusions in IoT networks reliably. The study runs binary and multi-class classification studies on the IoT Intrusion Detection dataset to identify a classifier that performs better than current techniques based on standard classification metrics. The research process follows the stages of the data mining lifecycle, which includes data collection and comprehension, data preparation, model building, validation, and assessment. In this context, six (6) to eight (8) well-established classification algorithms were tested on the IoT intrusion detection dataset to explore the posed research questions and hypotheses. The results of this investigation confirm the efficacy of ML techniques in identifying and classifying intrusion attempts from cyber-attacks on IoT networks, with precision and accuracy rates reaching 85%. The study underscores the significance of the IoT intrusion detection dataset as a critical tool for examining IoT network environments. It also reveals that decision tree ML algorithms are the most proficient classifiers tested in this study. In conclusion, the research provides an in-depth look at the factors that may influence the acceptance or rejection of each research hypothesis, offering valuable perspectives on the security of IoT networks.
- All rights reserved
Notice to Authors
If you are the author of this work and you have any questions about the information on this page, please use the Contact form to get in touch with us.