Detecting the Cyber Attacks on IoT-Based Network Devices Using Machine Learning Algorithms

Today, the number and variety of cyber-attacks on all systems have increased with the widespread use of internet technology. Within these systems, Internet of Things (IoT)-based network devices are especially exposed to a lot of cyber-attacks and are vulnerable to these attacks. This adversely affects the operation of the devices in question, and the data is endangered due to security vulnerabilities. Therefore, in this study, a model that detects cyber-attacks to ensure security with machine learning (ML) algorithms was proposed by using the data obtained from the log records of an IoT-based system. For this, first, the dataset was created, and this dataset was preprocessed and prepared by the models. Then, Artificial Neural Network (ANN), Random Forest (RF), K-Nearest Neighbor (KNN), Naive Bayes (NB), and Logistic Regression (LR) algorithms were used to create the models. As a result, the best performance to detect cyber-attacks was obtained using the RF algorithm with a rate of 99.6%. Finally, the results obtained from all the models created were compared with other academic studies in the literature and it was seen that the proposed RF model produced very successful results compared to the others. Moreover, this study showed that RF was a promising method of attack detection.