DC-STGCN: Dual-Channel Based Graph Convolutional Networks for Network Traffic Forecasting

Network traffic forecasting is essential for efficient network management and planning. Accurate long-term models are also proactive control of upcoming congestion events. Due to the complex spatial-temporal dependencies between flows, traditional time series often unable fully extract characteristics flows. To address this issue, we propose a novel dual-channel based graph convolutional (DC-STGCN) model. The proposed model consists two temporal components that characterize daily weekly correlation traffic. Each these contains extraction module consisting (DCGCN) gated recurrent unit (GRU). DCGCN further an adjacency feature (AGCN) (PGCN) capture connectivity nodes proximity correlation, respectively. GRU extracts experimental results on real data sets show prediction accuracy DC-STGCN overperforms existing baseline capable making predictions.