Traffic Classification Method Based on Graph Convolution Network

More and more derivative technologies are emerging in tandem with the Internet's ongoing growth, and these derivative technologies are also continuously improving current productivity. As a result, there is an increasing amount of complex and varied data traffic that the Internet must manage. This puts a greater strain on the network and makes it more difficult to classify network traffic, and it also raises the bar for the classification speed and accuracy of the classifier. The primary focus of this article is on the flaws in conventional network traffic classification algorithms. This study suggests an enhanced graph convolutional network traffic categorization method. This model includes an autoencoder and a graph convolutional network, and it can use the latter to extract. In order to fully describe the network traffic data, the self-encoder is also used to extract node features; these two parts of features are then combined; the combined features are then fed into the second layer of graph convolution; and finally, the Softmax classifier is used. Sort the second layer of graph convolution's output data into categories.