Application identification method for encrypted traffic of mobile terminal

The invention relates to a network traffic identification scene, and provides a mobile terminal encrypted traffic-oriented application identification method based on deep learning and an attention mechanism for mobile terminal applications and encrypted network traffic which are most widely applied at present. At the present stage, based on a deep learning model, traffic identification is performed in a parallel or serial mode by using different numbers of CNN and LSTM, an identification object is a foreign application data set for encrypted traffic, the identification object is not suitable for a domestic application environment, and the identification effect needs to be improved. A data set used in the method is collected and accurately marked for domestic common applications, and an attention mechanism is used for application identification on the basis of CNN and LSTM, so that domestic application identification can be effectively carried out, and the identification effect is better than that of a pure de