Machine Learning Algorithms for Binary and Multiclass Classification of IoT Network Traffic in Smart Cities

Internet of Things describes variety of embedded devices that are connected to the internet. This network of devices are used to build smart environment for smart city applications and used by many organizations. While data is constantly exchanged among various IoT devices, the detection of malicious attacks is essential for effective quality of service implementation in IoT networks. In this research, the University of New South Wales TON_IOT Train-Test dataset is used to classify network traffic. State of art machine learning methods such as Random Forest (RF), Decision Tree (DT), and Gradient Boosting (GB) are used to detect malicious IoT traffic. There are two classification types are considered in this study-binary and multi-label. When compared to the other two models, the Random Forest model gets the best accuracy of 99% in both binary and multi class classification.