Dynamic Path Flow Estimation Using Automatic Vehicle Identification and Probe Vehicle Trajectory Data: A 3D Convolutional Neural Network Model

Table 1

Performance of Model-A2 trained by samples with noisy labels.

	0			0.2			0.4
	0.2	0.4	0.6	0.2	0.4	0.6	0.2	0.4	0.6
RMSE veh/10 min	3.58	3.56	3.62	5.49	5.45	5.26	8.87	8.81	8.72
RMSE %	44.90	44.62	45.39	68.84	68.36	65.93	111.28	110.58	109.33
MAE veh/10 min	1.88	1.87	1.89	2.39	2.37	2.32	3.56	3.56	3.52
MAE %	23.54	23.48	23.76	29.95	29.73	29	44.69	44.6	44.12