Spatio-Temporal Hypergraph Convolutional Network Based Network Traffic Prediction

Abstract

Network traffic prediction plays an important role in network management and network operation and maintenance. Traditional network traffic prediction models do not take into account the impact of network routing paths on network performance, nor the impact of different traffic flows on the same link. To solve this challenge, this paper models the original network topology as a link hypergraph, taking the source-destination pairs of network traffic as the vertices in the graph and the links as hyperedges, which can effectively extract the connections between different application flows. influence relationship. On this basis, this paper proposes a network traffic prediction model based on spatio-temporal link hypergraph convolutional network, which can learn the temporal and spatial characteristics of network traffic at the same time. We conducted experiments on two real data sets. The experimental results show that the prediction model proposed in this paper can obtain the spatio-temporal correlation characteristics between network traffic and has higher prediction accuracy.