iguanas.rule_selection.CorrelatedFilter

class iguanas.rule_selection.CorrelatedFilter(correlation_reduction_class: iguanas.correlation_reduction.agglomerative_clustering_reducer.AgglomerativeClusteringReducer, rule_descriptions=None)[source]

Filters correlated rules based on a correlation reduction class (see the correlation_reduction sub-package).

Parameters
correlation_reduction_classAgglomerativeClusteringReducer

Instatiated class from the correlation_reduction sub-package.

rule_descriptionsPandasDataFrameType, optional

The standard performance metrics dataframe associated with the rules (if available). Defaults to None.

Attributes
rules_to_keepList[str]

List of rules which remain after correlated rules have been removed.

fit(X_rules: iguanas.utils.typing.pandas.core.frame.DataFrame, **kwargs) None[source]

Calculates the uncorrelated rules(using the correlation reduction class).

Parameters
X_rulesPandasDataFrameType

The binary columns of the rules applied to a dataset.

**kwargsdict

Any keyword arguments to pass to the correlation reduction class’s .fit() method

transform(X_rules: iguanas.utils.typing.pandas.core.frame.DataFrame) iguanas.utils.typing.pandas.core.frame.DataFrame[source]

Keeps only the uncorrelated rules in X_rules and rule_descriptions.

Parameters
X_rulesPandasDataFrameType

The binary columns of the rules applied to a dataset.

Returns
PandasDataFrameType

The binary columns of the uncorrelated rules.

fit_transform(X_rules: iguanas.utils.typing.pandas.core.frame.DataFrame, **kwargs) iguanas.utils.typing.pandas.core.frame.DataFrame[source]

Calculates the uncorrelated rules(using the correlation reduction class) then keeps only these uncorrelated rules in X_rules and rule_descriptions.

Parameters
X_rulesPandasDataFrameType

The binary columns of the rules applied to a dataset.

**kwargsdict

Any keyword arguments to pass to the correlation reduction class’s .fit() method.

Returns
PandasDataFrameType

The binary columns of the uncorrelated rules.