A machine-learning-based Intrusion detection for IIoT infrastructure

Industrial Internet of Things (IIoT) is rapidly expanding in the industrial sector, and it is critical to ensure the security of IIoT systems against cyber threats. This paper proposes an intrusion detection system for IIoT based on machine-learning algorithms. The proposed system uses a Decision Tree, Random Forest, and Naive Bayes algorithms to classify the incoming network traffic as normal or malicious. The dataset used in this research consists of 500,000 samples, including 250,000 attack samples and 250,000 normal samples. The dataset was preprocessed and features were extracted to train and test the proposed intrusion detection system. To evaluate the performance of the proposed system, accuracy, recall, precision, and F1-score, were used for evaluation. The results of the experiments show that the proposed system achieved high accuracy, recall, and precision in detecting malicious traffic in the IIoT network. The Decision Tree algorithm achieved the highest accuracy of 99.73%, followed by Random Forest with 99.70% accuracy and naive Bayes with 99.69% accuracy. The proposed intrusion detection system is expected to enhance the security of IIoT systems and mitigate the risk of cyber attacks in the industrial sector. It is recommended to integrate the proposed system with existing IIoT security systems to improve their effectiveness and provide a security solution.