Dropout regularization in Deep learning for Internet traffic classification

Nowadays, the Internet has become the global means of communication for transmitting all kinds of traffic data; this is what makes it develop rapidly, which means the network traffic data increases greatly. In order to manage the data circulating in the networks it is necessary to establish an analysis of the network traffic which makes it possible to provide detailed and explainable data on this traffic and the consumption of the bandwidth, generally by applications and their protocols. In this paper, we analyse and classify internet traffic using the « Deep Neurale Network » method and compare their performance to that of the artificial neural network. The results show that the proposed method is able to accurately and stably classify the content types of network traffic compared with to the RNN.