Hyperparameter Tuned Cloud Based Cyber Physical Attack Detection using Stacking Ensemble Learning
Rahul Sinha, Pragati Agrawal, Akhtar Rasool · 2024
The primary goal of this study is to analyze cyber-physical attacks on critical networks and to examine the operation of the cyber-physical system (intrusion detection system) and how hyperparameter-tuned ML models model are used to detect these attacks. Critical infrastructure, such as SCADA, WADI, and other remote control systems, require an advanced security solution to properly detect abnormalities and intrusions and safeguard data from malicious actors. The number of incursions in the modern era is rising significantly as a result of the fast expansion of linked devices, or "edge devices", which are open to assault. In this paper, we examine various machine learning techniques and apply stacking ensemble learning for our suggested approach. Next, we employ the hyperparameter tuning technique to enhance the model’s overall accuracy and performance. With our model, we have achieved 96.65% accuracy in the BATADAL dataset.