Cyber Attacks Classification on Enriching IoT Datasets

Abstract

In the era of the 5.0 industry, the use of the Internet of Things (IoT) has increased. The data generates from sensors through IoT industrial systems, any fault in those systems affects their performance and leads to real disaster. Protecting them from any possible attacks is an essential task. to secure any system, it needs to predict in the first place possible attacks and faults that could happen in the future. Predicting and initiating the attack type and the accuracy of these predictions can be done with machine learning models nowadays on the datasets produced with IoT networks. This paper classifies several attacks type based on several criteria and techniques to enhance the performance of machine learning (ML) models such as Voting techniques beside six ML models; Random Forest (RF), Decision Tree (DT), K-nearest neighbor (KNN), Support Vector Machine (SVM), Logistic regression (LR), and eXtreme Gradient Boosting (XGBoost) using Enriching IoT dataset. The results showed that 100% accuracy was achieved in estimating process with the XGBoost model.