An ensemble xgboost approach for the detection of cyber-attacks in the industrial iot domain

The Industrial Internet of Things (IIoT) connects numerous sensors, databases, machines, actuators, and working individuals. Our lives were improved by IIoT from multiple sectors including medical services, transportation, energy, agriculture, smart cities, associated vehicles, etc. Due to the rapid increase of IIoT devices in the industry, these systems are being frequently attacked by cyber-criminals. In this chapter, an ensemble XGBoost-based implementation for cyber-attack detection in the Industrial IoT is proposed. The presented technique is fast, lightweight, and it is assessed by utilizing the UNSW-NB15 dataset. The performance or execution of the presented technique is analyzed by various performance metrics like accuracy, F1 score, precision, log loss, and ROC. The analysis shows that the final accuracy of the proposed strategy is 96.88% with a learning rate of 0.05. The rest of the performance or evaluation parameters such as the precision, F1 score, recall, and AUC ROC score were 97.64%, 96.88%, 96.40%, and 96.88, respectively. It is also compared with various other algorithms such as multi-layer perceptron, artificial neural network, decision tree, logistic regression, and random forest, which prove the dominance of the proposed approach.