Rail Transit Prediction Based on Multi-View Graph Attention Networks

Table 3

Ablation contrast experiment.

Methods	July
	MAE	RMSE
Single view w/o autoencoder	9.21/10.31/12.38	19.68/22.49/23.51
Single view w/autoencoder	8.98/10.20/12.21	19.61/22.27/23.19
Multi-view w/o autoencoder	8.44/9.61/11.69	19.50/21.91/22.49
Multi-view w/autoencoder	8.35/9.57/11.60	19.41/21.84/22.38