Encrypted Traffic Classification Method Based on GSK and BiLSTM

As the global Internet moves towards full encryption, the significance of encrypted traffic classification is to provide an effective method to detect and monitor encrypted traffic in the network to detect and respond to potential security threats promptly. In addition, encrypted traffic classification can be used in areas such as network performance optimization, traffic management, and quality of service assurance. In this paper, we utilize the superior optimization-seeking capability of the GSK algorithm (human knowledge sharing algorithm) to optimize the initial learning rate, the maximum number of iterations, and the number of hidden layers of BiLSTM, and introduce an attention mechanism to assist in the classification of encrypted traffic. The model is classified using the Tor2016 dataset, and the experimental results show that the model classification accuracy reaches over 94%, showing high accuracy and robustness.