Network-Based Detection of Mirai Botnet Using Machine Learning and Feature Selection Methods
Ahmad Al–Qerem, Bushra Abutahoun, Shadi Ismail Nashwan, Shatha Shakhatreh, Mohammad Alauthman, Ammar Almomani · Advances in information security, privacy, and ethics book series · 2020
The spread of IoT devices is significantly increasing worldwide with a low design security that makes it more easily compromised than desktop computers. This gives rise to the phenomenon of IoT-based botnet attacks such as Mirai botnet, which have recently emerged as a high-profile threat that continues. Accurate and timely detection methods are required to identify these attacks and mitigate these new threats. To do so, this chapter will implement a network-based anomaly detection approach for the Mirai botnet using various machine learning and feature selection algorithms. Authors use Multiphase Genetic Algorithm section methods and PSO to select the best subfield of features capable of producing good overall classification results, and with this Feature Selection Algorithm, Random forest algorithm can detect all anomaly behavior with 100% accuracy.