BinningRecommendation#

class scikitplot.feature_selection.BinningRecommendation(feature, state, rule=None, n_bins=None, rule_n_bins=None, strategy=None, edges=(), reasons=())[source]#

A frozen bin recommendation for one feature.

Parameters:

featurestr

Column name.

statestr

Suitability, an AssessmentState value.

rulestr or None

Count rule used.

n_binsint or None

Recommended count (after clipping and quantile merging).

rule_n_binsint or None

Count the rule produced before clipping.

strategystr or None

"uniform" or "quantile".

edgestuple of float

n_bins + 1 increasing edges fitted on training data; empty when not suitable.

reasonstuple of str

Reason codes ("clipped_to_max_bins", "tied_quantiles_merged").

Parameters:

• feature (str)

• state (str)

• rule (str | None)

• n_bins (int | None)

• rule_n_bins (int | None)

• strategy (str | None)

• edges (tuple[float, ...])

• reasons (tuple[str, ...])

See also

BinningPolicy

The settings that produced this recommendation.

edges: tuple[float, ...] = ()#

feature: str[source]#

classmethod from_dict(data)[source]#

Rebuild from to_dict output, checking every field.

Parameters:

datamapping

Stored form.

Returns:

BinningRecommendation

The recommendation.

Raises:

ValueError

For an off-contract record.

Parameters:

data (Mapping[str, Any])

Return type:

BinningRecommendation

property is_suitable: bool#

The suitability state is "true".

Type:

bool

n_bins: int | None = None#

reasons: tuple[str, ...] = ()#

rule: str | None = None#

rule_n_bins: int | None = None#

state: str[source]#

strategy: str | None = None#

to_dict()[source]#

Return the shared-data-model form.

Return type:

dict[str, Any]