Research Article
Random Forest versus Support Vector Machine Models’ Applicability for Predicting Beam Shear Strength
Table 2
The prediction results for the investigated nine input combinations over the training and testing phases for the proposed RF model.
| Models combinations | R2 | RMSE (kN) | MAE (kN) | MAPE | Nash | MD |
| Training phase | M1 | 0.79 | 144.44 | 74.52 | 0.41 | 0.79 | 0.78 | M2 | 0.94 | 85.88 | 35.73 | 0.12 | 0.93 | 0.90 | M3 | 0.93 | 88.28 | 35.58 | 0.12 | 0.92 | 0.90 | M4 | 0.95 | 80.34 | | 0.09 | 0.93 | 0.92 | M5 | 0.95 | 79.58 | | 0.09 | 0.94 | 0.92 | M6 | 0.95 | 74.69 | 28.89 | 0.08 | 0.94 | 0.93 | M7 | 0.96 | 71.01 | 27.82 | 0.08 | 0.95 | 0.93 | M8 | 0.96 | 74.81 | 25.67 | 0.08 | 0.94 | 0.93 | M9 | 0.95 | 77.13 | 26.17 | 0.08 | 0.94 | 0.92 |
| Testing phase | M1 | 0.68 | 175.71 | 93.44 | 0.55 | 0.68 | 0.70 | M2 | 0.83 | 129.69 | 59.14 | 0.27 | 0.83 | 0.82 | M3 | 0.82 | 132.38 | 56.07 | 0.23 | 0.82 | 0.83 | M4 | 0.92 | 103.61 | 43.98 | 0.23 | 0.89 | 0.86 | M5 | 0.92 | 97.94 | 41.55 | 0.22 | 0.90 | 0.87 | M6 | 0.93 | 92.57 | 35.60 | 0.16 | 0.91 | 0.89 | M7 | 0.94 | 89.68 | 35.59 | 0.16 | 0.92 | 0.89 | M8 | 0.95 | 89.66 | 36.03 | 0.17 | 0.92 | 0.88 | M9 | 0.95 | 89.66 | 36.03 | 0.17 | 0.92 | 0.88 |
|
|