Network Traffic Classification via Deep Hashing Network

Network Traffic Classification via Deep Hashing Network

연구 분야: Networking

학회: International Artificial Intelligence Conference

The swift growth of the internet has resulted in a more complex and varied network traffic landscape, presenting substantial challenges for accurate traffic classification. Deep learning methods have provided innovative solutions for network traffic classification, but they often require model retraining to meet evolving classification demands. This paper introduces a new approach for classifying network traffic using deep hashing techniques, which maps network traffic data into binary hash codes for efficient classification. To the best of our knowledge, this is the first research to utilize deep hashing techniques for network traffic classification, bypassing the need for complex feature selection and utilizing raw traffic data as input. Experimental results using publicly available datasets indicate that our method attains high classification accuracy and performs exceptionally well in detecting malicious traffic.