Research Article
Road-Type Classification with Deep AutoEncoder
Table 4
Average micro
f1-score for DNN classifier across 10-fold CV at layers and
parameters.
| | | | Hidden layer size | Avg. micro f1-score | Hidden layer size | Avg. micro f1-score | Hidden layer size | Avg. micro f1-score |
| Layers A | 0.67 | Layers A | 0.77 | Layers A | 0.74 | Layers B | 0.71 | Layers B | 0.77 | Layers B | 0.69 | Layers C | 0.71 | Layers C | 0.79 | Layers C | 0.76 | Layers D | 0.74 | Layers D | 0.81 | Layers D | 0.78 |
|
|
Bold values indicate optimal DNN classifier parameters (hidden layers and learning rate) and the corresponding micro averaged f1-score.
|