Anonymous Traffic Detection Based on Feature Engineering and Reinforcement Learning

Area of Science:

• Computer Science
• Network Security
• Machine Learning

Background:

• Anonymous networks are crucial for user privacy but also facilitate malicious activities.
• Detecting anonymous network traffic is essential for internet security against evolving threats.
• Existing machine learning methods for detection often require extensive labeled datasets and complex models.

Purpose of the Study:

• To develop an efficient and accurate system for detecting anonymous network traffic.
• To overcome the limitations of existing methods, particularly the reliance on labeled data and complex architectures.
• To improve the training efficiency and performance of anonymous traffic detection systems.

Main Methods:

• Utilized feature engineering to extract and rank pattern information from static traces of anonymous traffic.
• Developed a reinforcement learning framework with states, actions, rewards, and state transitions.
• Implemented a lightweight system for classifying anonymous and non-anonymous traffic using fine-tuned thresholds instead of traditional labels.

Main Results:

• The proposed system successfully identifies anonymous network traffic.
• The system achieves an accuracy rate exceeding 80% based on pattern information.
• The approach demonstrates effective anonymous traffic detection without requiring labeled data.

Conclusions:

• The developed reinforcement learning framework offers a promising solution for detecting anonymous network traffic.
• The system's ability to perform detection without labeled data addresses a key challenge in the field.
• This research contributes to enhancing network security by providing a more efficient and accurate detection mechanism.