Table 4 Comparison of the performance of rut depth prediction models in different studies.
From: Applications of machine learning in predicting rut depth in off-road environments
References | Model used | Dataset size | Metrics (R2, RMSE) | Road type |
|---|---|---|---|---|
Linea regression | 70 | R2: 0.64 | On-road | |
ANN | 70 | R2: 0.87, RMSE: 1.49 | On-road | |
CatBoost | 1,630 | R2: 0.79 | On-road | |
XGBoost | 1,630 | R2: 0.73 | On-road | |
Random Forest (RF) | 1,630 | R2: 0.73 | On-road | |
LightGBM | 1,630 | R2: 0.69 | On-road | |
PSO-CatBoost | 1,630 | R2: 0.81 | On-road | |
PSO-XGBoost | 1,630 | R2: 0.80 | On- road | |
The current study | CatBoost | 270 | R2: 0.96, RMSE: 0.45 | Off-road |
The current study | GWO-CatBoost | 270 | R2: 0.97, RMSE: 0.38 | Off-road |
The current study | SBOA-CatBoost | 270 | R2: 0.97, RMSE: 0.35 | Off-road |
