Assessment#

class scikitplot.feature_selection.Assessment(name, state, provider, version, params=(), reasons=(), evidence=(), origin='inferred', inferred_state=None)[source]#

The recorded answer to one assessment for one feature.

Parameters:
namestr

Assessment name, as in the policy ("skewed").

statestr

Final AssessmentState value.

providerstr

Provider that answered.

versionstr

Provider version.

paramstuple of (str, value) pairs

Parameters used.

reasonstuple of str

Reason codes.

evidencetuple of (str, scalar) pairs

Evidence values.

origin{“inferred”, “manual”}, default=”inferred”

"manual" when a policy override set state.

inferred_statestr or None, default=None

For a manual override, the state the provider inferred; kept so the override never erases evidence.

Parameters:

See also

FeatureSpec.assess

Read an assessment of a fitted feature.

Examples

>>> import pandas as pd
>>> from scikitplot.feature_selection import profile_features
>>> schema = profile_features(pd.DataFrame({"x": [1.0, 1.0, 2.0, 50.0] * 5}))
>>> a = schema["x"].assess("skewed")
>>> a.state, a.reasons
('true', ('abs_skewness_ge_threshold',))
evidence: tuple[tuple[str, Any], ...] = ()#
property evidence_dict: dict[str, Any]#

Evidence as a mapping.

Type:

dict

classmethod from_dict(data)[source]#

Rebuild an assessment from to_dict output, checking every field.

Parameters:
datamapping

Stored form.

Returns:
Assessment

The assessment.

Raises:
ValueError

If a field is unknown, missing or wrongly typed.

Parameters:

data (Mapping[str, Any])

Return type:

Assessment

inferred_state: str | None = None#
property is_true: bool#

The state is "true" (unknown is not true).

Type:

bool

name: str[source]#
origin: str = 'inferred'#
params: tuple[tuple[str, Any], ...] = ()#
provider: str[source]#
reasons: tuple[str, ...] = ()#
state: str[source]#
to_dict()[source]#

Return the shared-data-model form.

Return type:

dict[str, Any]

version: str[source]#