Kuchi, Gopi (2024) Novel ways of Detecting threats in IIoT Networks. Masters thesis, Dublin, National College of Ireland.
Preview |
PDF (Master of Science)
Download (418kB) | Preview |
Preview |
PDF (Configuration Manual)
Download (913kB) | Preview |
Abstract
This study investigates how machine learning can improve cyber security in Industrial Internet of Things networks with an emphasis on devices that use the Modbus protocol. The study aims to a high level of threat detection accuracy using machine learning algorithms such as Random Forest Classifier and Support Vector Machine Classifier. The results in this experiment indicated that the ensembled-based Machine Learning Classifier has performed considerably better than Support Vector Machine. The accuracy for Random Forest Classifier is 98.64% and the accuracy for Support Vector Machine is 53.32%. The Runtime for Random Forest Classifier is 18.72 seconds, And Run time for Support Vector Machine is 38.22 seconds. In the future, work may focus on enhancing model interpretability and testing in real-world scenarios. All things considered, this research improves IIoT Cyber Security, By helping organizations find suitable algorithms that can analyze their traffic.
Actions (login required)
![]() |
View Item |