Deep Learning-Powered Multiclass Classification of DDoS Attacks on 6G-Connected IoT Devices

The rapid expansion of the 6G network and the widespread deployment of IoT devices resulted in challenging the effective detection and mitigation of distributed denial of service (DDoS) attacks originating from Internet of Things (IoT) sources. This paper presents an approach to handle this difficulty using machine learning and deep learning models. The approach uses Convolutional Neural Network (CNN) and Random Forest classifiers for binary classification, and Artificial Neural Network (ANN) model for determining the precise type of attack among nine classes. Fisher's Score and Recursive Feature Elimination with Cross Validation (RFECV) feature selection techniques are employed in the proposed approach for increasing the effectiveness of the system. The proposed approach is validated on the Canadian Institute for Cybersecurity-2019 dataset and the model achieves an accuracy rate of 99.5% for binary classification and more than 90% for different class classification.