iguanas.rule_application.RuleApplier

class iguanas.rule_application.RuleApplier(rule_strings: Dict[str, str], opt_func=None)[source]

Applies rules (stored in the standard Iguanas string format) to a dataset.

Parameters
rule_stringsDict[str, str]

Set of rules defined using the standard Iguanas string format (values) and their names (keys).

opt_funcCallable, optional

A function/method which calculates a custom metric (e.g. Fbeta score) for each rule. Defaults to None.

Attributes
rule_descriptionsPandasDataFrameType

Contains the logic of the rules and their performance metrics as applied to the dataset.

transform(X: Union[iguanas.utils.typing.pandas.core.frame.DataFrame, iguanas.utils.typing.databricks.koalas.frame.DataFrame], y=None, sample_weight=None) Union[iguanas.utils.typing.pandas.core.frame.DataFrame, iguanas.utils.typing.databricks.koalas.frame.DataFrame][source]

Applies the set of rules to a dataset, X. If y is provided, the performance metrics for each rule will also be calculated.

Parameters
XUnion[PandasDataFrameType, KoalasDataFrameType]

The feature set on which the rules should be applied.

yUnion[PandasSeriesType, KoalasSeriesType], optional

The target column. Defaults to None.

sample_weightUnion[PandasSeriesType, KoalasSeriesType], optional

Record-wise weights to apply. Defaults to None.

Returns
Union[PandasDataFrameType, KoalasDataFrameType]

The binary columns of the rules.