Adaptive Graph Convolutional Recurrent Network for Traffic Forecasting

Benchmark Model Rank Results
traffic-prediction-on-bjtaxiAGCRN#2MAE @ in: 12.30MAE @ out: 12.38MAPE (%) @ in: 15.61
traffic-prediction-on-expy-tky-1AGCRN#61 step MAE: 5.993 step MAE: 6.686 step MAE: 7.11
traffic-prediction-on-ne-bjAGCRN#612 steps MAE: 4.99
traffic-prediction-on-nycbike1AGCRN#2MAE @ in: 5.17MAE @ out: 5.47MAPE (%) @ in: 25.59
traffic-prediction-on-nycbike2AGCRN#2MAE @ in: 5.18MAE @ out: 4.79MAPE (%) @ in: 27.14
traffic-prediction-on-nyctaxiAGCRN#2MAE @ in: 12.13MAE @ out: 9.87MAPE (%) @ in: 18.78
traffic-prediction-on-pems04AGCRN#912 Steps MAE: 19.83
weather-forecasting-on-laAGCRN#2MSE (t+1): 0.2289 ± 0.0327MSE (t+6): 0.8412 ± 1.1162
weather-forecasting-on-noaa-atmosphericAGCRN#3MAE (t+1): 0.3019 ± 0.0374MAE (t+10): 1.3755 ± 0.2732
weather-forecasting-on-sdAGCRN#2MSE (t+1): 0.2010 ± 0.0188MSE (t+6): 1.0181 ± 0.1275