ClassificationMetric

class ClassificationMetric[source]

Bases: Metric, ABC

A base class for classification metrics.

Attributes Summary

`binarize`	whether the metric needs binarized scores
`closed_expectation`	whether there is a closed-form solution of the expectation
`closed_variance`	whether there is a closed-form solution of the variance
`key`	Return the key for use in metric result dictionaries.
`supports_weights`	whether the metric supports weights
`synonyms`	synonyms for this metric

Methods Summary

`__call__`(y_true, y_score[, weights])	Evaluate the metric.
`extra_repr`()	Generate the extra repr, cf.
`forward`(y_true, y_score[, sample_weight])	Calculate the metric.
`get_description`()	Get the description.
`get_link`()	Get the link from the docdata.
`get_range`()	Get the math notation for the range of this metric.
`iter_extra_repr`()	Iterate over the components of the `extra_repr()`.

Attributes Documentation

binarize: ClassVar[bool | None] = None: whether the metric needs binarized scores

closed_expectation: ClassVar[bool] = False: whether there is a closed-form solution of the expectation

closed_variance: ClassVar[bool] = False: whether there is a closed-form solution of the variance

key: Return the key for use in metric result dictionaries.

supports_weights: ClassVar[bool] = False: whether the metric supports weights

synonyms: ClassVar[Collection[str]] = (): synonyms for this metric

Methods Documentation

__call__(y_true: ndarray, y_score: ndarray, weights: ndarray | None = None) → float[source]

Evaluate the metric.

Parameters:

y_true (ndarray) – shape: (num_samples,) the true labels, either 0 or 1.
y_score (ndarray) – shape: (num_samples,) the predictions, either continuous or binarized.
weights (ndarray | None) –
shape: (num_samples,) weights for individual predictions

Warning

not all metrics support sample weights - check supports_weights first

Returns:

the scalar metric value

Raises:

ValueError – when weights are provided but the function does not support them.

Return type:

float

extra_repr() → str

Generate the extra repr, cf. :meth`torch.nn.Module.extra_repr`.

Returns:: the extra part of the repr()
Return type:: str

abstractmethod forward(y_true: ndarray, y_score: ndarray, sample_weight: ndarray | None = None) → float[source]

Calculate the metric.

Parameters:

y_true (ndarray) – shape: (num_samples,) the true label, either 0 or 1.
y_score (ndarray) – shape: (num_samples,) the predictions, either as continuous scores, or as binarized prediction (depending on the concrete metric at hand).
sample_weight (ndarray | None) – shape: (num_samples,) sample weights

Returns:

a scalar metric value

Return type:

float

# noqa:DAR202

classmethod get_description() → str

Get the description.

Return type:: str

classmethod get_link() → str

Get the link from the docdata.

Return type:: str

classmethod get_range() → str

Get the math notation for the range of this metric.

Return type:: str

iter_extra_repr() → Iterable[str]

Iterate over the components of the extra_repr().

This method is typically overridden. A common pattern would be

def iter_extra_repr(self) -> Iterable[str]:
    yield from super().iter_extra_repr()
    yield "<key1>=<value1>"
    yield "<key2>=<value2>"

Returns:: an iterable over individual components of the extra_repr()
Return type:: Iterable[str]