<ul><li>Study introduces a spatiotemporal modeling approach combining Graph Convolutional Networks (GCN) with Gated Recurrent Units (GRU) for predicting network traffic in complex topological environments.</li><li>GCN captures spatial dependencies among network nodes, while GRU models the temporal evolution of traffic data, enabling precise forecasting of future traffic patterns.</li><li>Proposed model validated through experiments on real-world Abilene network traffic dataset, outperforming popular deep learning methods in performance metrics.</li><li>Ablation experiments conducted to evaluate components such as graph convolution layers, temporal modeling strategies, and adjacency matrix construction, showing superior performance and strong generalization capabilities.</li></ul>

Joint Graph Convolution and Sequential Modeling for Scalable Network Traffic Estimation

Discover more