Which element of the documentation do you want to modify and why?
In the documentation of julearn.run_cross_validation, it states that only ScorerLike is accepted as a scoring parameter, but a list of strings, for example, ["roc_auc", "balanced_accuracy"], is also admitted. I think it should be clearer that multiple metrics can be used in the scoring, as this is a big feature of the library.
It would also be good to clearly state that this scoring is not used for hyperparameter tuning, and that it is a separate parameter.
Finally, shouldn't None be stated as the default parameter?
Anything else to say?
Current documentation:
scoring: ScorerLike, optional
The scoring metric to use. See https://scikit-learn.org/stable/modules/model_evaluation.html for a comprehensive list of options. If None, use the model’s default scorer.
Which element of the documentation do you want to modify and why?
In the documentation of
julearn.run_cross_validation, it states that only ScorerLike is accepted as ascoringparameter, but a list of strings, for example, ["roc_auc", "balanced_accuracy"], is also admitted. I think it should be clearer that multiple metrics can be used in the scoring, as this is a big feature of the library.It would also be good to clearly state that this scoring is not used for hyperparameter tuning, and that it is a separate parameter.
Finally, shouldn't
Nonebe stated as the default parameter?Anything else to say?
Current documentation:
scoring: ScorerLike, optionalThe scoring metric to use. See https://scikit-learn.org/stable/modules/model_evaluation.html for a comprehensive list of options. If None, use the model’s default scorer.