Deep Learning Model under Complex Network and its Application in Traffic Detection and Analysis

This paper proposes a deep learning model based on convolutional neural networks for the current complex network environment to classify network traffic. The load of each network traffic is converted into a two-dimensional gray image, and the generated image is used as the input of the model. Convert network flow classification problems into image classification problems. The purpose of network flow classification is achieved by classifying images. The payload data of network traffic is a continuous one-dimensional byte stream organized in a hierarchical structure. The bytes, data packets, and conversations in the payload correspond exactly to the characters, words, and sentences in the field of natural language processing. Therefore, the load of network traffic can be regarded as a sentence, and then these "sentences" as input, through the model classification processing, to complete the classification of network traffic. In this paper, the deep learning model is applied to network traffic classification research. The model automatically learns relevant features from traffic data. It can liberate the researchers from the heavy feature learning and feature selection work, which has certain advantages in classification accuracy compared with traditional methods.