To figure out if the performance (e.g. in terms of the RMSE) significantly differ between models, a non-parametric, two-sided Mann-Whitney U-test will be integrated.
The MW U-test will be part of the meta-postprocessing to ensure that the required scores are already available.
For implementation, the mannwhitneyu-method of the scipy-package can be used as also done in MLAir.
To figure out if the performance (e.g. in terms of the RMSE) significantly differ between models, a non-parametric, two-sided Mann-Whitney U-test will be integrated. The MW U-test will be part of the meta-postprocessing to ensure that the required scores are already available.
For implementation, the mannwhitneyu-method of the scipy-package can be used as also done in MLAir.