Detection and Characterization of Darknet Traffic Using Attention LSTM with XAI
G. Kirubavathi, Y Amithesh · 2024
The Darknet comprises networks and technologies prioritizing privacy and security, often linked to illicit activities involving malware and attacks on legitimate services. To combat potential Darknet misuse, it is crucial to comprehend and classify its traffic. This study delves into analyzing authentic Darknet traffic using the CIC-Darknet2020 dataset. By conducting meticulous data pre-processing, an Attention Long Short-Term Memory (LSTM) model is trained to identify and categorize Darknet traffic accurately. The results are remarkable, with a detection accuracy of 99.70% for Darknet traffic and 97% for classifying the nature of Darknet operations. Furthermore, the LIME technique offers transparent explanations for the model's decisions. Compared to cutting-edge methodologies, this proposed system surpasses others, showcasing the approach's efficacy in detecting and comprehending Darknet traffic.