Short-Term Speed Prediction Using Remote Microwave Sensor Data: Machine Learning versus Statistical Model

Table 3

Prediction accuracy of models for different forecasting steps ahead in station C.

MAE	Number of forecasting steps ahead
	1	3	5	10
BPNN	2.5694	3.4613	4.1736	5.6407
NARXNN	2.5484	3.4757	4.1458	5.7245
SVM-RBF	2.6617	3.6111	4.2583	5.6255
SVM-LIN	2.8896	3.7087	4.3468	5.6221
MLR	3.9650	3.8713	4.5163	5.9062
ARIMA	2.6675	3.6235	4.2857	5.6745
VAR	2.6611	3.6798	4.3051	5.4380
ST	2.6359	3.5298	4.0963	5.2024
MAPE (%)	Number of forecasting steps ahead
	1	3	5	10
BPNN	6.5606	8.9994	11.3564	16.3149
NARXNN	6.4875	9.1176	11.2402	16.6231
SVM-RBF	6.7398	9.3483	11.5343	16.4376
SVM-LIN	7.0707	10.1794	10.9789	15.5072
MLR	7.1717	10.4244	11.3601	16.7207
ARIMA	6.7822	9.4230	11.6734	16.7901
VAR	6.6898	9.1819	11.0211	14.5708
ST	6.6709	8.9993	10.7423	14.3994