Adaptive Security in a Connected World: Review of Machine Learning for IoT Intrusion Detection

In IoT Security risks are, however, presented to devices and services by this integration. To give machine learning (ML) credit for being a novel approach, this paper surveys Intrusion Detection Systems (IDS) for the Traditional Network and Internet of Things. With 41 articles review, Machine Learning can adapt to the changing IoT environment, Deep Learning (DL) in particular shows promise as a method. IoT device with security problem with complex IoT landscape is too much for traditional rule-based IDS to handle, which has led to a rise in interest in machine learning approaches. To solve issues used different techniques are investigated. The research discusses important feature selection techniques related to IoT intrusion detection, emphasizing the value of feature engineering and classification. With newly created datasets both normal and malicious scenarios use performance measures such as F1-score, accuracy, precision, recall, and AUC-ROC.