Table 5.
Measurement of model performance for GAI-prediction [m2 m-2] with advanced predictive models in calibration and evaluation with ratios of reflections as predictors.
| Advanced Predictive Models | MAEcalibration (RMSEcalibration) | MAEevaluation (RMSEevaluation) | Equation | |
|---|---|---|---|---|
| Partial Least Squares | 0.36 (0.55) | 0.47 (0.76) | … | (7) |
| Support Vector Machine (linear Kernel) | 0.36 (0.54) | 0.44 (0.71) | … | (8) |
| Support Vector Machine (radial Kernel) | 0.32 (0.51) | 0.44 (0.69) | … | (9) |
| K Nearest Neighbor | 0.33 (0.54) | 0.47 (0.72) | … | (10) |
| Multivariate Adaptive Regression Spline | 0.32 (0.51) | 0.44 (0.70) | … | (11) |
| Boosted Trees | 0.29 (0.47) | 0.52 (0.82) | … | (12) |
| mean | 0.33 (0.52) | 0.46 (0.73) |
MAEs are colored dark grey if they are higher than the mean and white if equal or lower.