A Bayesian Rule Learning Based Intrusion Detection System for the MQTT Communication Protocol

What is it about?

Our work differs from previous works in following aspects: Firstly, our study not only implements various attack scenarios for data generation but also investigates a concrete attack detection method. Secondly, in our experiment all devices are physically installed, which creates a very realistic test environment, as it can be seen in section 3. Thirdly, our work goes beyond the DoS attack detection, which is commonly seen in intrusion detection research. Fourthly, for intrusion detectionwe consider not only network layer features but also application layer features. Lastly, by using rule learning method we aim not only for good detection performance but also for high interpretability.

Featured Image

Photo by S. Tsuchiya on Unsplash

Photo by S. Tsuchiya on Unsplash