LabelRankerMixin

class sklr.base.LabelRankerMixin

Mixin class for all label rankers in scikit-lr.

A label ranker predicts a ranking of the labels for each sample, without ties, and learns from (possibly incomplete) rankings without ties. This mixin sets the estimator_type tag to "label_ranker" and states through the target tags that fit requires a target y with several outputs, one per label, as described in Ranking representation. Its score() method computes Kendall’s \(\tau\) between the predicted and the true rankings.

See also

PartialLabelRankerMixin

Mixin class for all partial label rankers.

is_label_ranker

Return whether an estimator is a label ranker.

Examples

>>> import numpy as np
>>> from sklearn.base import BaseEstimator
>>> from sklr.base import LabelRankerMixin, is_label_ranker
>>> # Mixin classes should always be on the left-hand side for a correct MRO
>>> class MyEstimator(LabelRankerMixin, BaseEstimator):
...     def fit(self, X, y):
...         self.ranking_ = np.asarray(y)[0]
...         return self
...     def predict(self, X):
...         return np.tile(self.ranking_, (len(X), 1))
>>> estimator = MyEstimator()
>>> X = np.array([[1, 2], [2, 3], [3, 4]])
>>> y = np.array([[1, 2, 3], [2, 1, 3], [1, 3, 2]])
>>> estimator.fit(X, y).predict(X)
array([[1, 2, 3],
       [1, 2, 3],
       [1, 2, 3]])
>>> is_label_ranker(estimator)
True

score(X: ArrayLike, y: ArrayLike, sample_weight: ArrayLike | None = None) → float

Return the mean Kendall’s \(\tau\) on the given data.

It is computed with kendall_tau_score().

Parameters:

Xarray-like of shape (n_samples, n_features)

Test samples.

yarray-like of shape (n_samples, n_labels)

True rankings for X, complete and without ties.

sample_weightarray-like of shape (n_samples,), default=None

Sample weights.

Returns:

scorefloat

Mean Kendall’s \(\tau\) of self.predict(X) with respect to y.