Research Article
Machine Learning for Estimating Leaf Dust Retention Based on Hyperspectral Measurements
Table 2
Training and testing results for three plants using the different regression algorithms.
| Plants | Algorithms | Number of bands | Training | Testing | | MAE (g/m2) | RMSE (g/m2) | Bias | Number of samples | | MAE (g/m2) | RMSE (g/m2) | Bias | Number of samples |
| GL | CART | 1000 | 0.98 | 0.04 | 0.08 | 0.00 | 99 | 0.78 | 0.17 | 0.26 | 0.05 | 42 | 94 | 0.95 | 0.08 | 0.13 | 0.00 | 0.76 | 0.18 | 0.26 | 0.04 | RF | 1000 | 0.98 | 0.07 | 0.10 | 0.00 | 0.88 | 0.14 | 0.19 | 0.05 | 94 | 0.98 | 0.06 | 0.08 | 0.00 | 0.92 | 0.11 | 0.15 | 0.03 | SVM | 1000 | 0.97 | 0.08 | 0.09 | −0.01 | 0.89 | 0.14 | 0.19 | 0.05 | 94 | 0.92 | 0.13 | 0.16 | 0.00 | 0.89 | 0.16 | 0.19 | 0.07 | LC | CART | 1000 | 0.97 | 0.05 | 0.10 | 0.00 | 97 | 0.50 | 0.33 | 0.47 | −0.06 | 41 | 167 | 0.95 | 0.07 | 0.12 | 0.00 | 0.50 | 0.29 | 0.46 | −0.03 | RF | 1000 | 0.97 | 0.08 | 0.11 | 0.00 | 0.83 | 0.17 | 0.23 | −0.03 | 167 | 0.98 | 0.07 | 0.10 | 0.00 | 0.87 | 0.14 | 0.20 | −0.02 | SVM | 1000 | 0.97 | 0.09 | 0.10 | −0.01 | 0.80 | 0.17 | 0.23 | −0.02 | 167 | 0.95 | 0.11 | 0.13 | 0.00 | 0.77 | 0.20 | 0.25 | −0.01 | CF | CART | 1000 | 0.93 | 0.08 | 0.13 | 0.00 | 94 | 0.64 | 0.20 | 0.31 | 0.00 | 41 | 73 | 0.98 | 0.04 | 0.07 | 0.00 | 0.70 | 0.17 | 0.29 | 0.03 | RF | 1000 | 0.97 | 0.06 | 0.09 | 0.00 | 0.80 | 0.16 | 0.23 | 0.02 | 73 | 0.98 | 0.05 | 0.08 | 0.00 | 0.84 | 0.14 | 0.21 | 0.02 | SVM | 1000 | 0.97 | 0.06 | 0.08 | 0.01 | 0.90 | 0.13 | 0.17 | 0.03 | 73 | 0.97 | 0.08 | 0.08 | 0.00 | 0.88 | 0.14 | 0.18 | −0.01 |
|
|