Table 4 Performance of soil free iron content models based on principal component analysis (PCA) spectrum.
From: Estimation of soil free Iron content using spectral reflectance and machine learning algorithms
Spectral | Modeling | Training | Testing | ||||
|---|---|---|---|---|---|---|---|
Transform | Methods | R2 | RMSE (g/kg) | RRMSE (%) | R2 | RMSE (g/kg) | RRMSE (%) |
Original | PLS | 0.641 | 7.02 | 31.2 | 0.609 | 7.041 | 32.6 |
SVM | 0.705 | 6.633 | 31.4 | 0.645 | 6.881 | 31.4 | |
RF | 0.953 | 3.244 | 14.7 | 0.736 | 6.334 | 28.9 | |
DNN | 0.805 | 5.453 | 24.6 | 0.751 | 5.902 | 26.9 | |
FD | PLS | 0.847 | 4.579 | 20.6 | 0.76 | 5.539 | 25.7 |
SVM | 0.876 | 4.085 | 18.8 | 0.803 | 5.231 | 23.8 | |
RF | 0.99 | 4.497 | 20.3 | 0.460 | 10.325 | 47.0 | |
DNN | 0.900 | 4.597 | 19.4 | 0.807 | 5.875 | 26.8 | |
SNV | PLS | 0.642 | 7.019 | 31.4 | 0.625 | 6.967 | 32.3 |
SVM | 0.713 | 6.408 | 30.3 | 0.623 | 7.264 | 33.1 | |
RF | 0.950 | 3.396 | 15.4 | 0.639 | 7.204 | 32.8 | |
DNN | 0.812 | 5.203 | 23.7 | 0.752 | 5.955 | 27.1 | |
CR | PLS | 0.790 | 7.019 | 31.4 | 0.732 | 6.967 | 32.3 |
SVM | 0.726 | 6.408 | 30.3 | 0.715 | 7.264 | 33.1 | |
RF | 0.960 | 3.396 | 15.4 | 0.703 | 7.204 | 32.8 | |
DNN | 0.826 | 5.203 | 23.7 | 0.715 | 5.955 | 27.1 | |
