Comparative Analysis of Cyber Security Approaches Using Machine Learning in Industry 4.0

Abstract

The concept of Industry 4.0 emerged due to the increase in production in the industry, the need for high efficiency in production and the desire to benefit from the features of technology. In Industry 4.0, the data from the systems can be used in real time with the internet technology, the operation of the systems is automated using the internet technology. can be remotely controlled and managed by system administrators. In addition to the contribution of technology to the industrial field, network technology that enables critical systems to communicate with each other causes systems to be exposed to cyber attacks. In this study, it is mentioned that signature-based attack detection systems aren't effective in new generation attacks and machine learning algorithms are successful in detecting next generation attacks. In order to verify, the data obtained from the traffic of the industrial system network and machine learning (ML) algorithms were trained and the performances of the algorithms in predicting cyber attacks were compared. The performance of each algorithm shows their accuracy in predicting attacks. Traditional Decision Tree, Random Forest, K Nearest Neighbor (KNN), Naive Bayes, Logistic Regression and Support Vector Machine (SVM) algorithms are used to detect attacks. With this study, it is aimed to gain a perspective on attack detection studies with machine learning techniques in Industry 4.0.