Archives of Advances in Artificial Intelligence & Data Science and Machine Learning | Volume 1, Issue 1 | Review Article | Open Access
Soren Falkner*
Vienna University of Technology, Faculty of Computer Engineering, Vienna, Austria
*Correspondence to: Soren Falkner
Fulltext PDFThe proliferation of Internet of Things (IoT) devices has introduced unprecedented connectivity and data generation, but also significant security challenges. Traditional security mechanisms often prove inadequate for the dynamic and resource-constrained nature of IoT networks. Anomaly detection, leveraging the power of machine learning, offers a promising approach to identify unusual or malicious behavior within these complex environments. This paper explores the application of various machine learning techniques, including supervised, unsupervised and semi-supervised methods, for detecting anomalies in IoT data streams. We discuss the unique characteristics of IoT data that influence the choice and performance of these algorithms, such as high dimensionality, temporal dependencies and the prevalence of noisy or imbalanced datasets. Furthermore, we examine the challenges and opportunities associated with deploying machine learning-based anomaly detection systems in resource-constrained IoT environments, including model training, real-time inference and data privacy considerations. Finally, we highlight promising research directions and potential advancements in this critical area for securing the future of connected devices.
Anomaly detection; Internet of things (IoT); Machine learning; Security; Intrusion detection; Unsupervised learning; Supervised learning; Semi-supervised learning; Time series analysis; Edge computing; Resource constraints; Data security; Cyber security
Falkner S. Anomaly Detection in IoT Using Machine Learning. Arch Adv Art Intel Data Sci Mach Learn 2025;1(1):1-11.