Table 20 ML models’ performance metrics [26]
From: A review of optimization techniques and algorithms used for FRP applications in civil engineering
Model | Training dataset | Test dataset | ||||||
|---|---|---|---|---|---|---|---|---|
RMSE (kN.m) | MAE (kN.m) | MAPE (%) | R2(%) | RMSE (kN.m) | MAE (kN.m) | MAPE (%) | R2(%) | |
KNN | 4.80 | 3.10 | 5.87 | 97.2 | 8.43 | 5.15 | 9.30 | 93.1 |
KRR | 4.45 | 3.56 | 7.86 | 97.6 | 4.71 | 3.37 | 7.32 | 97.8 |
SVR | 3.97 | 2.72 | 5.77 | 98.1 | 4.44 | 3.04 | 6.01 | 98.1 |
CART | 3.95 | 2.58 | 5.86 | 98.1 | 5.14 | 3.68 | 8.44 | 97.4 |
RF | 3.71 | 2.67 | 5.52 | 98.4 | 4.30 | 3.51 | 7.92 | 98.2 |
GBT | 2.52 | 1.84 | 3.91 | 99.2 | 3.19 | 2.33 | 4.73 | 99.0 |
xgBoost | 2.41 | 1.55 | 3.17 | 99.3 | 2.70 | 1.77 | 3.25 | 99.3 |