Efficient Traffic Classification Using Hybrid Deep Learning

Network traffic classification is important for QoS, security, and billing. Using a single machine learning classification model suffers from low prediction and classification accuracy, especially for high dimensional datasets with a high sparsity level. Recently, hybrid-deep learning algorithms have shown high efficiency for traffic forecasting and classification. In this research, we suggested two hybrid models – the Convolutional Neural Network (CNN) combined with the Recurrent Neural Network (RNN) models and the Gated recurrent unit (GRU) combined with Long ShortTerm Memory (LSTM) – to improve traffic classification accuracy in terms of the ratio of correct predictions to total input samples. The hybrid CNN-LSTM and CNN-GRU have achieved an accuracy of up to 99.23% and 93.92%, respectively, for binary classification and 67.16% for multiclass classification. We employed Markov Blanket Algorithm to make the reduced feature set selection, which significantly improved classification accuracy and fscore for multi-class scenarios.