A Comparative Analysis of Supervised Machine Learning Models for Smart Intrusion Detection in IoT Network
Saksham Mittal, Amit Kumar Mishra, Vikas Tripathi, Prabhdeep Singh, Priyank Pandey
Abstract
IoT will become an integral part of everyone's life in making the world more intelligent and smarter. In an IoT ecosystem, objects collect and share information in order to communicate with one another. The objective of this paper is to design a safe and secure IoT network, which is more effective and smart in intrusion detection by identifying potential threats and different attacks and initiating an appropriate action. To build such a smart system, complex machine learning models are required which are trained on intrusion detection datasets. In this paper, at first two different benchmark datasets: CICIDS2017 and Bot-IoT are described briefly, which can be used for a smart intrusion detection system. After this, there is a discussion about the use of four different mathematical models using supervised ML algorithms on both the datasets respectively to classify different types of attacks, followed by a comparative analysis of the results obtained from the models in the results section.