An image classification network for network traffic representation learning

• Published in: 2021 6th International Symposium on Computer and Information Processing Technology (ISCIPT) pp. 358 - 361
• Main Author: Liu, Xinyan
• Format: Conference Proceeding
• Language: English
• Published: IEEE 01.06.2021
• Subjects: Classification algorithms; deep learning; image classification; Internet; MobileNet; Network security; network traffic; Planning; Representation learning; Telecommunication traffic; Training
• DOI: 10.1109/ISCIPT53667.2021.00078

Network traffic classification builds a mapping between traffic data and application types according to traffic characteristics, which is one of the basic tasks in the field of network planning, maintenance management, and network security[5]. Due to the rapid development of network technology and the rapid increase of network traffic, it is extremly necessary and urgent for rapid and accurate automatic classification of network traffic. Through the research of ResNet[3], DenseNet[4], GoogleN et[1] and other classic convolutional neural networks, this paper proposes a lightweight network traffic classification algorithm, which completes the network traffic classification task based on the design ideas and structural advantages of the MobileNet[2]. The comparative experimental results demonstrate that the algorithm can significantly reduce the training time and model size, and achieves a superior performance of network traffic classification.