Traffic Classification Based on Graph Convolutional Network
Xingguo Ji, Qingmin Meng · 2020
Traffic classification is the first step of network QoS control mechanism and traffic anomaly detection and is an important research branch of congestion control and network security. In complex network analysis, the traffic classification scheme based on machine learning often requires a large number of correctly labeled samples, which requires a lot of manual operations. Therefore, it is still a tough thing for traffic classification with higher accuracy at low identification rates. In this work, two network representation based on graph convolution network (GCN) are tried to avoid a large number of labeled samples. The research method is convenient to combine the representation learning based on the graph structure network with the network traffic problems, and obtain a certain classification accuracy of network traffic with relatively few marking samples. The preliminary experimental results show that the proposed method achieves good performance under small network scale, and the classification accuracy rate reaches 97.35%.