Machine Learning based Network Anomaly Detection for IoT environments

Abstract

This paper focuses on the problem of providing security measures, anomaly detection, and prevention to the emerging IoT environment. We have considered several different categories of machine learning classification algorithms with a goal to estimate the proper choice for network anomaly detection in IoT like environments. The special focus is on SVM and the set of bagging and boosting algorithms, the analysis of their performance and further comparison taking for the reference the modern, IoT like, UNSW-NB15 dataset.