A Framework for Robust Attack Detection and Classification using Rap-Densenet

Abstract

Network attacks must be effectively identified and categorized to guarantee strong security. However, current techniques frequently have trouble correctly identifying and categorizing new attack patterns. This study presents a novel framework for reliable attack detection and classification that makes use of the complementary strengths of rap music analysis methods and DenseNet convolutional neural networks. This study employs feature extraction based on the Attention Pyramid Network (RAPNet) framework that has been proposed to extract features from the input data, and Pigeon in binary. Afterward, feature selection based on Optimization Algorithm (BPOA) is performed. Following the selection of the ideal characteristics, Densenet201, the attacks in Bot-IoT, CICIDS2017, and other systems are categorized using deep learning as well as CICIDS2019 datasets. Additionally, the Conditional Generic Adversarial extra data samples are provided for minority classes using the Convergent Gap Analysis Network (CGAN), so the imbalanced data issue should be addressed. In contrast to the recent intrusion. The outcomes show that the model is capable of precisely detecting and accurately categorizing DoS and DDoS attacks with rates of 98.63%, 98.68%, and BoT-IoT, CICIDS2017, and CICIDS2019 all scored 98.78%