From: ASENN: attention-based selective embedding neural networks for road distress prediction
Method | RMSE | \(R^2\) | MAE | MAPE |
---|---|---|---|---|
Random Forest [5] | 0.017 | 0.999 | 0.005 | 0.012 |
XGBoost [6] | 0.010 | 0.999 | 0.003 | 0.010 |
LightGBM [7] | 0.021 | 0.999 | 0.007 | 0.017 |
CatBoost [8] | 0.024 | 0.999 | 0.007 | 0.029 |
MLP [3] | 0.110 | 0.986 | 0.027 | 0.147 |
ResNet [14] | 0.038 | 0.998 | 0.025 | 0.213 |
FT-transformer [10] | 0.009 | 0.999 | 0.006 | 0.022 |
TabNet [9] | 0.060 | 0.995 | 0.044 | 0.182 |
Ours | 0.002 | 0.999 | 0.001 | 0.006 |