An Efficient and Smart Botnet Attack Detection Model for IIoT 4.0

The Industrial Internet of Things (IIoT) is a rapidly expanding and pervasive network of networked objects, including industrial sensors, actuators, and other IoT devices. It is becoming a new rule for the future. Because of the complex IIOT environment, any attack on the system severely affects the overall performance. Unfortunately, there have been a lot of botnet attacks in IIoT network which calls for security. A botnet attack detection system helps to locate, aim, and neutralize botnet attacks on networks. This paper focuses on building a machine learning (ML) model to categorize botnet attacks and explores diverse ML methods. The objective is to differentiate harmful botnet activity from regular network behavior automatically. Staying ahead of emerging threats requires integrating various technologies and ensuring regular updates. Using the UNSW-NB 15 dataset, this study trains many algorithms (Decision Tree, XGBoost, Logistic Regression, K-Nearest Neighbor, and Random Forest) to detect Botnet attacks. F1 Score, Recall, Accuracy, and Precision are used to gauge performance. KNN beats all other models and even the state-of-the-art models, achieving the maximum accuracy of 99%.