Table 4 Estimated performance indices for the studied ML models in the training stage.
From: Machine learning and interactive GUI for concrete compressive strength prediction
Type | Model | R2 | WI | RMSE (MPa) | SI | MAE (MPa) | MAPE | MBE (MPa) |
|---|---|---|---|---|---|---|---|---|
Non-ensemble | MLR | 0.617 | 0.870 | 10.34 | 0.151 | 8.23 | 0.313 | 0.132 |
MNLR | 0.567 | 0.836 | 11.00 | 0.186 | 8.54 | 0.357 | − 0.319 | |
SVR | 0.597 | 0.867 | 10.61 | 0.138 | 8.11 | 0.310 | − 0.56 | |
GEP | 0.913 | 0.946 | 7.21 | 0.090 | 5.66 | 0.201 | − 0.072 | |
ANN* | 0.972 | 0.993 | 2.77 | 0.036 | 1.17 | 0.042 | − 0.194 | |
ANFIS | 0.871 | 0.965 | 6.00 | 0.073 | 4.61 | 0.172 | − 0.001 | |
Ensemble | RF | 0.985 | 0.996 | 2.05 | 0.027 | 1.32 | 0.046 | 0.077 |
AdaBoost | 0.835 | 0.946 | 6.80 | 0.126 | 5.63 | 0.258 | − 1.100 | |
XGBoost | 0.986 | 0.996 | 2.00 | 0.026 | 1.38 | 0.050 | − 0.012 | |
CatBoost* | 0.987 | 0.997 | 1.93 | 0.025 | 1.33 | 0.048 | − 0.001 |
