STGNN-FAM: A Traffic Flow Prediction Model for Spatiotemporal Graph Networks Based on Fusion of Attention Mechanisms

Table 3

Comparison of prediction performance of different models on PeMSD7 dataset.

Model	15 min			30 min			45 min
	MAE	RMSE	MAPE	MAE	RMSE	MAPE	MAE	RMSE	MAPE
HA	4.01	7.20	10.61	4.01	7.20	10.61	4.01	7.20	10.61
ARIMA	5.55	9.00	12.92	5.86	9.13	13.94	6.27	9.38	15.20
LSVR	2.50	4.55	5.81	3.63	6.67	8.88	4.54	8.28	11.50
FNN	2.74	4.75	6.38	4.02	6.98	9.72	5.04	8.58	12.38
FC-LSTM	3.57	6.20	8.60	3.94	7.03	9.55	4.16	7.51	10.10
DCRNN	2.37	4.21	5.54	3.31	5.96	8.06	4.01	7.13	9.99
STGCN	2.25	4.04	5.26	3.03	5.70	7.33	3.57	6.77	8.69
ST-GAT	2.01	3.45	4.76	2.76	4.68	6.57	3.20	5.30	7.86
STGNN-FAM	1.98	3.50	4.84	2.55	4.51	6.26	2.90	5.17	7.10

The bold values in Table 3 represent the best performance.