Research Article
Short-Term Speed Prediction Using Remote Microwave Sensor Data: Machine Learning versus Statistical Model
Table 2
Prediction accuracy of models for different forecasting steps ahead in station B.
| MAE | Number of forecasting steps ahead | 1 | 3 | 5 | 10 |
| BPNN | 2.9402 | 3.1513 | 3.3712 | 3.8135 | NARXNN | 2.9297 | 3.1434 | 3.3591 | 3.8208 | SVM-RBF | 2.9825 | 3.2323 | 3.4342 | 3.8567 | SVM-LIN | 3.1945 | 3.2111 | 3.5712 | 3.8806 | MLR | 3.2624 | 3.9019 | 4.1148 | 4.5207 | ARIMA | 3.0144 | 3.2669 | 3.4683 | 3.8959 | VAR | 3.0244 | 3.3033 | 3.7154 | 4.0356 | ST | 3.4509 | 3.2993 | 3.5093 | 3.9322 |
| MAPE (%) | Number of forecasting steps ahead | 1 | 3 | 5 | 10 |
| BPNN | 4.9518 | 5.3793 | 5.7737 | 6.6007 | NARXNN | 4.9325 | 5.3470 | 5.7459 | 6.6096 | SVM-RBF | 4.9811 | 5.4445 | 5.8253 | 6.6268 | SVM-LIN | 5.4039 | 5.4243 | 6.2740 | 6.8330 | MLR | 5.5909 | 6.4957 | 6.8970 | 7.6689 | ARIMA | 5.0383 | 5.5078 | 5.8899 | 6.7010 | VAR | 5.0470 | 5.6490 | 6.4091 | 6.8736 | ST | 5.7732 | 5.5666 | 5.9707 | 6.7795 |
|
|