Highway Travel Time Prediction of Segments Based on ANPR Data considering Traffic Diversion

Table 8

Short-term prediction error results of different methods.

Methods	Error	5 min	10 min	15 min
XGBoost	RMSE	11.1040	12.2234	12.9625
	MAE	3.0112	3.3359	3.4920
	MAPE (%)	14.6106	16.2274	16.7007
LR	RMSE	10.9548	11.9806	12.6909
	MAE	3.7284	4.3713	4.7821
	MAPE (%)	22.1522	25.7278	27.8272
SARIMA	RMSE	10.0307	11.2200	11.9780
	MAE	2.6366	3.1975	3.5113
	MAPE (%)	15.7659	18.9068	20.3997
CNNs	RMSE	11.8657	12.4348	12.8615
	MAE	5.8780	5.7638	5.6719
	MAPE (%)	35.5655	34.3175	33.2884
LSTM-ST	RMSE	13.8093	13.9326	14.0234
	MAE	4.5960	4.6382	4.6554
	MAPE (%)	23.0800	23.3385	23.4371
LSTM-CNN	RMSE	8.4288	9.7757	10.7455
	MAE	1.7900	2.3881	2.7997
	MAPE (%)	10.3360	13.1795	14.9992