LabelAmbiguity

class LabelAmbiguity

Find samples with multiple labels.

Parameters

columnsUnion[Hashable, List[Hashable]] , default: None

List of columns to check, if none given checks all columns Except ignored ones.

ignore_columnsUnion[Hashable, List[Hashable]] , default: None

List of columns to ignore, if none given checks based on columns variable.

n_to_showint , default: 5

number of most common ambiguous samples to show.

__init__(columns: Optional[Union[Hashable, List[Hashable]]] = None, ignore_columns: Optional[Union[Hashable, List[Hashable]]] = None, n_to_show: int = 5)

__new__(*args, **kwargs)

Methods

LabelAmbiguity.add_condition(name, ...)	Add new condition function to the check.
LabelAmbiguity.add_condition_ambiguous_sample_ratio_not_greater_than([...])	Add condition - require samples with multiple labels to not be more than max_ratio.
LabelAmbiguity.clean_conditions()	Remove all conditions from this check instance.
LabelAmbiguity.conditions_decision(result)	Run conditions on given result.
LabelAmbiguity.name()	Name of class in split camel case.
LabelAmbiguity.params([show_defaults])	Return parameters to show when printing the check.
LabelAmbiguity.remove_condition(index)	Remove given condition by index.
LabelAmbiguity.run(dataset[, model])	Run check.
LabelAmbiguity.run_logic(context[, dataset_type])	Run check.

Example

LabelAmbiguity Example Notebook

previous

NewLabelTrainTest.run_logic

next

LabelAmbiguity.add_condition