Using Machine Learning in Detecting IOT Cyber Attacks

Abstract

Abstract: The Internet of Things (IoT) combines hundreds of millions of devices that are able to interact with each other with minimal user interaction. IoT is one of the fastest growing areas of computing; however, the reality is that in the extremely hostile environment of the Internet, IoT is vulnerable to many types of cyberattacks. Practical countermeasures to secure IoT networks, such as network anomaly detection, need to be implemented to address this issue. Regardless of the fact that attacks cannot be completely avoided forever, early detection of an attack is essential for practical defense. As IoT devices have low storage capacity and low computing power, traditional high-end security solutions are not suitable for IoT system protection. IoT devices are also now connected without human intervention for longer periods of time. This means that intelligent networkbased security solutions such as machine learning solutions need to be developed. Although many studies in recent years have discussed the application of Machine Learning (ML) solutions to attack detection problems, little attention has been paid to attack detection specifically in IoT networks. In this study, we aim to contribute to the literature by evaluating various machine learning algorithms that can be used to quickly and effectively detect IoT network attacks. A new Bot-IoT dataset is used to evaluate different detection algorithms. Seven different machine learning algorithms were used in the implementation phase and most of them achieved high performance. New features were extracted from the Bot-IoT dataset during implementation and compared with literature studies, and the new features provided better results