Machine Learning Based IoT IntrusionDetection System: An MQTT Case Study

This work uses six different machine learning techniques to classify attacks in an MQTT network.

Dataset Used

The used dataset is published in IEEE DataPort

@data{bhxy-ep04-20,
doi = {10.21227/bhxy-ep04},
url = {http://dx.doi.org/10.21227/bhxy-ep04},
author = {Hanan Hindy; Christos Tachtatzis; Robert Atkinson; Ethan Bayne; Xavier Bellekens },
publisher = {IEEE Dataport},
title = {MQTT Internet of Things Intrusion Detection Dataset},
year = {2020} } 

Citation

@article{hindy2020machine,
  title={Machine Learning Based IoT Intrusion Detection System: An MQTT Case Study},
  author={Hindy, Hanan and Bayne, Ethan and Bures, Miroslav and Atkinson, Robert and Tachtatzis, Christos and Bellekens, Xavier},
  journal={arXiv preprint arXiv:2006.15340},
  year={2020}
}

Algorithms Used

• Logistic Regression
• k-Nearest Neighbours
• Gaussian Naive Bayes
• Decision Trees
• Random Forests
• Support Vector Machine (linear and RBF kernel)

How to Run it:

Clone this repository
Download dataset files and extract them in the same directory
run classification.py --mode [0: packet, 1: unidirectional, 2: bidirectional] --output [output_folder] --verbose [True/False]

• The classification outputs are added to the output folder.