Research Article

A Dynamic Spatio-Temporal Deep Learning Model for Lane-Level Traffic Prediction

Table 2

The performance of our model and other baselines on traffic flow prediction. Bold scores are the best and underlined scores are the second best.

T5 min10 min15 min

ModelMAERMSEVARMAERMSEVARMAERMSEVAR

MLP3.45474.36160.53053.40334.46730.49023.74874.78430.4248
LSTM3.21214.14080.56873.30414.22250.55263.31714.23760.5494
GRU3.12504.12290.56593.31864.23680.55273.32974.25110.5479
GCN4.80865.99250.08364.80285.98630.08554.80525.97990.0876
T-GCN4.44675.46640.25414.57565.60310.21864.52715.63920.1879
Ours w/o gate2.83113.77300.63802.92493.85930.62213.01613.94910.6049
Ours2.82743.76970.63852.91683.85220.62342.99513.92680.6093