Rail Transit Prediction Based on Multi-View Graph Attention Networks

Figure 1

The end-to-end framework of proposed model. The MV-GAT model includes multi-view input, multi-layer attention module, and the autoencoder module. The node-level prediction results are obtained as output.