Research Article
Coupling a Bat Algorithm with XGBoost to Estimate Reference Evapotranspiration in the Arid and Semiarid Regions of China
Table 6
Results of the statistical indicators for all models in the Taiyuan station during the training and testing stages.
| Model | Training | Testing | R2 | RMSE | MAE | RMSEmax10% | MAEmax10% | R2 | RMSE | MAE | RMSEmax10% | MAEmax10% |
| GPR1 | 0.812 | 0.605 | 0.466 | 1.34 | 1.311 | 0.817 | 0.714 | 0.564 | 1.457 | 1.451 | GPR2 | 0.972 | 0.248 | 0.19 | 0.561 | 0.55 | 0.96 | 0.418 | 0.311 | 0.949 | 0.942 | GPR3 | 0.97 | 0.244 | 0.18 | 0.594 | 0.564 | 0.889 | 0.534 | 0.317 | 1.487 | 1.328 | GPR4 | 0.972 | 0.243 | 0.176 | 0.582 | 0.564 | 0.98 | 0.403 | 0.31 | 0.835 | 0.827 | GPR5 | 0.976 | 0.223 | 0.16 | 0.545 | 0.534 | 0.955 | 0.391 | 0.297 | 0.85 | 0.842 | GPR6 | 0.978 | 0.22 | 0.152 | 0.559 | 0.546 | 0.976 | 0.397 | 0.301 | 0.858 | 0.851 | MARS1 | 0.949 | 0.317 | 0.24 | 0.712 | 0.703 | 0.935 | 0.497 | 0.351 | 1.138 | 1.115 | MARS2 | 0.97 | 0.241 | 0.187 | 0.542 | 0.534 | 0.976 | 0.377 | 0.275 | 0.865 | 0.853 | MARS3 | 0.974 | 0.224 | 0.163 | 0.551 | 0.52 | 0.96 | 0.377 | 0.259 | 0.954 | 0.884 | MARS4 | 0.984 | 0.178 | 0.132 | 0.422 | 0.406 | 0.986 | 0.364 | 0.284 | 0.811 | 0.803 | MARS5 | 0.982 | 0.189 | 0.137 | 0.453 | 0.438 | 0.958 | 0.385 | 0.282 | 0.868 | 0.859 | MARS6 | 0.988 | 0.151 | 0.101 | 0.39 | 0.361 | 0.978 | 0.33 | 0.22 | 0.817 | 0.809 | BAXGB1 | 0.972 | 0.241 | 0.181 | 0.555 | 0.542 | 0.939 | 0.412 | 0.302 | 0.928 | 0.918 | BAXGB2 | 0.994 | 0.115 | 0.087 | 0.266 | 0.256 | 0.97 | 0.314 | 0.24 | 0.712 | 0.699 | BAXGB3 | 0.984 | 0.227 | 0.16 | 0.556 | 0.532 | 0.976 | 0.259 | 0.193 | 0.581 | 0.58 | BAXGB4 | 0.996 | 0.097 | 0.071 | 0.232 | 0.221 | 0.986 | 0.297 | 0.235 | 0.643 | 0.632 | BAXGB5 | 0.996 | 0.082 | 0.063 | 0.186 | 0.182 | 0.964 | 0.308 | 0.237 | 0.722 | 0.715 | BAXGB6 | 1 | 0.036 | 0.026 | 0.083 | 0.08 | 0.988 | 0.253 | 0.182 | 0.594 | 0.578 |
|
|
Note. Bold font indicates the best model.
|