Abstract

In cybersecurity, identifying and reducing cyber threats within encrypted Virtual Private Network (VPN) and The Onion Router (Tor) traffic data pose noteworthy obstacles. To tackle this issue, we present a new and innovative deep learning methodology incorporating Grey Wolf Optimization (GWO) to perform feature selection. This approach aims to enhance the effectiveness and precision of cyber threat detection models by optimizing their efficiency. Our research aims to effectively identify and address cyber risks by utilizing Recurrent Neural Networks (RNNs) to capture temporal dependencies in the data, which are crucial for this purpose. The efficacy of the proposed model is thoroughly assessed using the CIC-Darknet2020 dataset, and the findings substantiate the dominance of the GWO-driven RNN model, attaining a remarkable accuracy rate of 99.312%. This study highlights the efficacy of the Grey Wolf Optimization (GWO) approach in enhancing cybersecurity measures and strengthening network security against encrypted communication threats.