check_scoring#

sklearn.metrics.check_scoring(estimator=None, scoring=None, *, allow_none=False, raise_exc=True)[source]#

Determine scorer from user options.

A TypeError will be thrown if the estimator cannot be scored.

Parameters:

estimatorestimator object implementing ‘fit’ or None, default=None

The object to use to fit the data. If None, then this function may error depending on allow_none.

scoringstr, callable, list, tuple, set, or dict, default=None

Scorer to use. If scoring represents a single score, one can use:

a single string (see The scoring parameter: defining model evaluation rules);
a callable (see Defining your scoring strategy from metric functions) that returns a single value.

If scoring represents multiple scores, one can use:

a list, tuple or set of unique strings;
a callable returning a dictionary where the keys are the metric names and the values are the metric scorers;
a dictionary with metric names as keys and callables a values. The callables need to have the signature callable(estimator, X, y).

If None, the provided estimator object’s score method is used.

allow_nonebool, default=False

Whether to return None or raise an error if no scoring is specified and the estimator has no score method.

raise_excbool, default=True

Whether to raise an exception (if a subset of the scorers in multimetric scoring fails) or to return an error code.

If set to True, raises the failing scorer’s exception.
If set to False, a formatted string of the exception details is passed as result of the failing scorer(s).

This applies if scoring is list, tuple, set, or dict. Ignored if scoring is a str or a callable.

Added in version 1.6.

Returns:

scoringcallable: A scorer callable object / function with signature scorer(estimator, X, y).

Examples

>>> from sklearn.datasets import load_iris
>>> from sklearn.metrics import check_scoring
>>> from sklearn.tree import DecisionTreeClassifier
>>> X, y = load_iris(return_X_y=True)
>>> classifier = DecisionTreeClassifier(max_depth=2).fit(X, y)
>>> scorer = check_scoring(classifier, scoring='accuracy')
>>> scorer(classifier, X, y)
0.96...

>>> from sklearn.metrics import make_scorer, accuracy_score, mean_squared_log_error
>>> X, y = load_iris(return_X_y=True)
>>> y *= -1
>>> clf = DecisionTreeClassifier().fit(X, y)
>>> scoring = {
...     "accuracy": make_scorer(accuracy_score),
...     "mean_squared_log_error": make_scorer(mean_squared_log_error),
... }
>>> scoring_call = check_scoring(estimator=clf, scoring=scoring, raise_exc=False)
>>> scores = scoring_call(clf, X, y)
>>> scores
{'accuracy': 1.0, 'mean_squared_log_error': 'Traceback ...'}