An Open Access Journal
Table 4 Training time for different types of models
From: Models for forecasting the traffic flow within the city of Ljubljana
Model type | Train time [s] | Inference time [s] |
|---|---|---|
Mean | 0.19 | 0.01 |
Regressor model | 125.68 | 0.16 |
Decision tree (None, 2) | 23.79 | 0.06 |
Extremely randomized trees (None, 2) | 145.51 | 0.57 |
Neural network (128, 1) | 229.69 | 0.30 |
Neural network (256, 1) | 236.93 | 0.31 |
Neural network (512, 1) | 266.45 | 0.32 |
Neural network (128, 2) | 655.69 | 0.52 |
Neural network (256, 2) | 655.69 | 0.57 |
Neural network (512, 2) | 642.70 | 0.59 |
Neural network (128, 3) | 819.35 | 0.74 |
Neural network (256, 3) | 820.70 | 0.87 |
Neural network (512, 3) | 899.02 | 1.86 |
Neural network (128, 4) | 930.53 | 1.47 |
Neural network (256, 4) | 1080.80 | 2.08 |
Neural network (512, 4) | 1293.85 | 2.64 |
Neural network (128, 5) | 1263.38 | 1.54 |
Neural network (256, 5) | 1486.36 | 2.45 |
Neural network (512, 5) | 1510.61 | 3.40 |