From: ASENN: attention-based selective embedding neural networks for road distress prediction
Method | RMSE | \(R^2\) | MAE | MAPE |
---|---|---|---|---|
Random forest [5] | 0.782 | 0.401 | 0.236 | 0.694 |
XGBoost [6] | 0.771 | 0.417 | 0.233 | 0.687 |
LightGBM [7] | 0.763 | 0.430 | 0.247 | 0.726 |
CatBoost [8] | 0.760 | 0.433 | 0.228 | 0.627 |
MLP [3] | 0.758 | 0.437 | 0.233 | 0.681 |
ResNet [14] | 0.762 | 0.431 | 0.243 | 0.757 |
FT-transformer [10] | 0.759 | 0.435 | 0.227 | 0.626 |
TabNet [9] | 0.766 | 0.425 | 0.250 | 0.650 |
Ours | 0.728 | 0.481 | 0.218 | 0.573 |