Smartsentry Cyber Threat Intelligence in IOT

Abstract: The SmartSentry system has established itself as the Cyber Threat Intelligence (CTI) system in the Industrial Internet of Things (IIoT) area owing to its possibility to identify threats right down to the source of the data. To do this, additional detectors are applied along with more sophisticated machine learning and deep learning algorithms that carry out comprehensive and detailed analyses of the cyber threats. A team of researchers has put together an entire collection of algorithms to single out the threat and to determine its severity based on the type of information, which is composed of the following: Decision Tree (DT), Extra Trees Classifier (ETC), Support Vector machine (SVM), k-Nearest Neighbor (KNN), and Deep Neural Network (DNN). As a part of our IIoT anomaly detection study, we handled the issue of data imbalance with the integration of the Synthetic Minority Over-sampling Technique (SMOTE) into our method. The Decision Tree and Random Forest models got accuracy scores of 0.9979, while the SVM model was at 0.7644, the Extra Trees model at 0.9961, and KNN at 0.9196. The SmartSentry's capability to find and stop anomalies, which is the latter of the two cyber threat tactics - the proactive one thereby ensuring IIoT systems stability and safety, will be a great leap forward in the war to protect IIoT. Keywords: IIoT, Cyber Threat Intelligence, Machine Learning, Deep Learning, Anomaly Detection, Random Forest, Decision Tree, Support Vector Machine, k-Nearest Neighbor