Feature mining for encrypted malicious traffic detection with deep learning and other machine learning algorithms

The popularity of encryption mechanisms poses a great challenge to malicious traffic detection. reason is traditional detection techniques cannot work without the decryption encrypted traffic. Currently, research on has focused feature extraction and choice machine learning or deep algorithms. In this paper, we first provide an in-depth analysis features compare different state-of-the-art creation approaches, while proposing novel concept for which specifically designed analysis. addition, propose framework two-layer consists both Through comparative experiments, it outperforms classical algorithms, such as ResNet Random Forest. Moreover, sufficient training data model, also curate dataset composed entirely public datasets. more comprehensive than using any alone. Lastly, discuss future directions research.