Effective Classification of Tor Traffic Using Deep Learning
摘要
The growing prevalence of online privacy concerns has propelled the use of the Tor network as a popular means for secure and anonymous communication. Tor, “The Onion Router,” is a privacy-focused network that aims to provide anonymity and security for users browsing the Internet. Tor works by routing Internet traffic, encrypting the data at every step. Tor network is based on various attributes such as protocol, content, destination, and behavior. We have used machine learning and deep learning for Tor classification, where feature learning is done using Convolutional Neural Network. Classification of Tor traffic can be done by extracting the information from the Tor traffic network. Our work has achieved 99.77 accuracy for Random forest Model and 96.60 accuracy for Deep Neural Network model which has outperformed the best work to date.