Transfer learning-based Mirai botnet detection in IoT networks
Sana Rabhi, Tarek Abbes, Faouzi Zarai · 2023
The Internet of Things (IoT) has expeditiously evolved from a recent concept to a prevalent and often crucial side of diverse environments, including industry, healthcare, homes and buildings, etc. Its applications for almost all the fields have actually become a principal target for various threats, mainly botnets. To mitigate this recent threat, diverse techniques were developed, particularly those founded on Machine Learning, Deep Learning and Transfer Learning which have achieved significant results in detecting IoT botnets. In our paper we propose a transfer learning intrusion detection system (IDS) to identify IoT botnets, comparing between deep learning and transfer learning models using two different algorithms, Artificial Neural Network (ANN) and Recurrent Neural Network (RNN). Results prove that the suggested approach based TL reaches high accuracy up to 99% that overcomes the accuracy of DL model for the two algorithms.