causalnex.discretiser.DecisionTreeSupervisedDiscretiserMethod¶

class causalnex.discretiser.DecisionTreeSupervisedDiscretiserMethod(mode='single', split_unselected_feat=False, tree_params=None)[source]¶

Bases: causalnex.discretiser.abstract_discretiser.AbstractSupervisedDiscretiserMethod

Allows the discretisation of continuous features based on the split thresholds of either sklearn’s DecisionTreeRegressor or DecisionTreeClassifier. DecisionTreeSupervisedDiscretiserMethod is inhereited from AbstractSupervisedDiscretiserMethod. When instantiated, we have an object with .fit method to learn discretisation thresholds from data and .transform method to process the input.

Example:

 import pandas as pd
 import numpy as np
 from causalnex.discretiser.discretiser_strategy import DecisionTreeSupervisedDiscretiserMethod
 from sklearn.datasets import load_iris
 iris = load_iris()
 X, y = iris["data"], iris["target"]
 names = iris["feature_names"]
 data = pd.DataFrame(X, columns=names)
 data["target"] = y
 dt_multi = DecisionTreeSupervisedDiscretiserMethod(
     mode="multi", tree_params={"max_depth": 3, "random_state": 2020}
 )
 tree_discretiser = dt_multi.fit(
     feat_names=[
         "sepal length (cm)",
         "sepal width (cm)",
         "petal length (cm)",
         "petal width (cm)",
     ],
     dataframe=data,
     target="target",
     target_continuous=False,
 )
 discretised_data = tree_discretiser.transform(data[["petal width (cm)"]])
 discretised_data.values.ravel()
array([0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
   0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
   0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
   1, 1, 1, 1, 2, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
   1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 2, 2, 2, 2, 2, 2, 1, 2, 2, 2,
   2, 2, 2, 2, 2, 2, 2, 2, 2, 1, 2, 2, 2, 2, 2, 2, 2, 2, 2, 1, 2, 2,
   2, 1, 1, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2])

Attributes

Methods

`DecisionTreeSupervisedDiscretiserMethod.__delattr__`(name, /)	Implement delattr(self, name).
`DecisionTreeSupervisedDiscretiserMethod.__dir__`()	Default dir() implementation.
`DecisionTreeSupervisedDiscretiserMethod.__eq__`(…)	Return self==value.
`DecisionTreeSupervisedDiscretiserMethod.__format__`(…)	Default object formatter.
`DecisionTreeSupervisedDiscretiserMethod.__ge__`(…)	Return self>=value.
`DecisionTreeSupervisedDiscretiserMethod.__getattribute__`(name, /)	Return getattr(self, name).
`DecisionTreeSupervisedDiscretiserMethod.__getstate__`()
`DecisionTreeSupervisedDiscretiserMethod.__gt__`(…)	Return self>value.
`DecisionTreeSupervisedDiscretiserMethod.__hash__`()	Return hash(self).
`DecisionTreeSupervisedDiscretiserMethod.__init__`([…])	This Discretiser Method uses Decision Trees to predict the target.
`DecisionTreeSupervisedDiscretiserMethod.__init_subclass__`(…)	Set the `set_{method}_request` methods.
`DecisionTreeSupervisedDiscretiserMethod.__le__`(…)	Return self<=value.
`DecisionTreeSupervisedDiscretiserMethod.__lt__`(…)	Return self<value.
`DecisionTreeSupervisedDiscretiserMethod.__ne__`(…)	Return self!=value.
`DecisionTreeSupervisedDiscretiserMethod.__new__`(…)	Create and return a new object.
`DecisionTreeSupervisedDiscretiserMethod.__reduce__`()	Helper for pickle.
`DecisionTreeSupervisedDiscretiserMethod.__reduce_ex__`(…)	Helper for pickle.
`DecisionTreeSupervisedDiscretiserMethod.__repr__`([…])	Return repr(self).
`DecisionTreeSupervisedDiscretiserMethod.__setattr__`(…)	Implement setattr(self, name, value).
`DecisionTreeSupervisedDiscretiserMethod.__setstate__`(state)
`DecisionTreeSupervisedDiscretiserMethod.__sizeof__`()	Size of object in memory, in bytes.
`DecisionTreeSupervisedDiscretiserMethod.__sklearn_clone__`()
`DecisionTreeSupervisedDiscretiserMethod.__str__`()	Return str(self).
`DecisionTreeSupervisedDiscretiserMethod.__subclasshook__`	Abstract classes can override this to customize issubclass().
`DecisionTreeSupervisedDiscretiserMethod._build_request_for_signature`(…)	Build the MethodMetadataRequest for a method using its signature.
`DecisionTreeSupervisedDiscretiserMethod._check_feature_names`(X, …)	Set or check the feature_names_in_ attribute.
`DecisionTreeSupervisedDiscretiserMethod._check_n_features`(X, …)	Set the n_features_in_ attribute, or check against it.
`DecisionTreeSupervisedDiscretiserMethod._get_default_requests`()	Collect default request values.
`DecisionTreeSupervisedDiscretiserMethod._get_metadata_request`()	Get requested data properties.
`DecisionTreeSupervisedDiscretiserMethod._get_param_names`()	Get parameter names for the estimator
`DecisionTreeSupervisedDiscretiserMethod._get_tags`()
`DecisionTreeSupervisedDiscretiserMethod._more_tags`()
`DecisionTreeSupervisedDiscretiserMethod._repr_html_inner`()	This function is returned by the @property _repr_html_ to make hasattr(estimator, “_repr_html_”) return `True or False depending on get_config()[“display”].
`DecisionTreeSupervisedDiscretiserMethod._repr_mimebundle_`(…)	Mime bundle used by jupyter kernels to display estimator
`DecisionTreeSupervisedDiscretiserMethod._transform_one_column`(…)	Given one “original” feature (continuous), discretise it.
`DecisionTreeSupervisedDiscretiserMethod._validate_data`([…])	Validate input data and set or check the n_features_in_ attribute.
`DecisionTreeSupervisedDiscretiserMethod._validate_params`()	Validate types and values of constructor parameters
`DecisionTreeSupervisedDiscretiserMethod.fit`(…)	The fit method allows DecisionTrees to learn split thresholds from the input data
`DecisionTreeSupervisedDiscretiserMethod.fit_transform`(…)	raises NotImplementedError fit_transform is not implemented
`DecisionTreeSupervisedDiscretiserMethod.get_metadata_routing`()	Get metadata routing of this object.
`DecisionTreeSupervisedDiscretiserMethod.get_params`([deep])	Get parameters for this estimator.
`DecisionTreeSupervisedDiscretiserMethod.set_fit_request`(*)	Request metadata passed to the `fit` method.
`DecisionTreeSupervisedDiscretiserMethod.set_params`(…)	Set the parameters of this estimator.
`DecisionTreeSupervisedDiscretiserMethod.set_transform_request`(*)	Request metadata passed to the `transform` method.
`DecisionTreeSupervisedDiscretiserMethod.transform`(data)	Given one “original” dataframe, discretise it.

__init__(mode='single', split_unselected_feat=False, tree_params=None)[source]¶

This Discretiser Method uses Decision Trees to predict the target. The cutting points on the the Decision Tree becomes the chosen discretisation thresholds

If the target is a continuous variable, we fit a DecisionTreeRegressor to discretise the data. Otherwise, we fit a Classifier.

Parameters

max_depth (int) – maximum depth of the decision tree.
mode (str) – Either ‘single’ or ‘multi’.
if single (-) – The splitting points of the decision tree become discretiser fixed points
a univariate decision tree for each continuous variable being discretised. (Train) – The splitting points of the decision tree become discretiser fixed points
if multi (-) – The splitting points of each variable used in the Decision tree become the thresholds for discretisation
a decision tree over all the variables passed. (Train) – The splitting points of each variable used in the Decision tree become the thresholds for discretisation
split_unselected_feat (bool) – only applicable if self.mode = ‘multi’.
if True (-) –
not selected by the decision tree will be discretised using 'single' mode (features) –
the same tree parameters (with) –
if False (-) –
not selected by the decision tree will be left unchanged (features) –
tree_params (Optional[Dict[str, Any]]) – keyword arguments, which are parameters
for sklearn.tree.DecisionTreeClassifier/sklearn.tree.DecisionTreeRegressor (used) –

Raises

KeyError – if an incorrect argument is passed

fit(feat_names, target, dataframe, target_continuous)[source]¶

The fit method allows DecisionTrees to learn split thresholds from the input data

Parameters

feat_names (List[str]) – a list of feature to be discretised
target (str) – name of variable that is going to be used a target for the decision tree
dataframe (pd.DataFrame) – pandas dataframe of input data
target_continuous (bool) – a boolean that indicates if the target variable is continuous

Returns

DecisionTreeSupervisedDiscretiserMethod object with learned split thresholds from the decision tree

Return type

self

fit_transform(*args, **kwargs)¶

Raises: NotImplementedError – fit_transform is not implemented

get_metadata_routing()¶

Get metadata routing of this object.

Please check User Guide on how the routing mechanism works.

Returns: routing – A MetadataRequest encapsulating routing information.
Return type: MetadataRequest

get_params(deep=True)¶

Get parameters for this estimator.

Parameters: deep (bool, default=True) – If True, will return the parameters for this estimator and contained subobjects that are estimators.
Returns: params – Parameter names mapped to their values.
Return type: dict

set_fit_request(*, dataframe: Union[bool, None, str] = '$UNCHANGED$', feat_names: Union[bool, None, str] = '$UNCHANGED$', target: Union[bool, None, str] = '$UNCHANGED$', target_continuous: Union[bool, None, str] = '$UNCHANGED$') → causalnex.discretiser.discretiser_strategy.DecisionTreeSupervisedDiscretiserMethod¶

Request metadata passed to the fit method.

Note that this method is only relevant if enable_metadata_routing=True (see sklearn.set_config()). Please see User Guide on how the routing mechanism works.

The options for each parameter are:

True: metadata is requested, and passed to fit if provided. The request is ignored if metadata is not provided.
False: metadata is not requested and the meta-estimator will not pass it to fit.
None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.
str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

New in version 1.3.

Note

This method is only relevant if this estimator is used as a sub-estimator of a meta-estimator, e.g. used inside a pipeline.Pipeline. Otherwise it has no effect.

Parameters

dataframe (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for dataframe parameter in fit.
feat_names (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for feat_names parameter in fit.
target (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for target parameter in fit.
target_continuous (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for target_continuous parameter in fit.

Returns

self – The updated object.

Return type

object

set_params(**params)¶

Set the parameters of this estimator.

The method works on simple estimators as well as on nested objects (such as Pipeline). The latter have parameters of the form <component>__<parameter> so that it’s possible to update each component of a nested object.

Parameters: **params (dict) – Estimator parameters.
Returns: self – Estimator instance.
Return type: estimator instance

set_transform_request(*, data: Union[bool, None, str] = '$UNCHANGED$') → causalnex.discretiser.discretiser_strategy.DecisionTreeSupervisedDiscretiserMethod¶

Request metadata passed to the transform method.

Note that this method is only relevant if enable_metadata_routing=True (see sklearn.set_config()). Please see User Guide on how the routing mechanism works.

The options for each parameter are:

True: metadata is requested, and passed to transform if provided. The request is ignored if metadata is not provided.
False: metadata is not requested and the meta-estimator will not pass it to transform.
None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.
str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

New in version 1.3.

Note

This method is only relevant if this estimator is used as a sub-estimator of a meta-estimator, e.g. used inside a pipeline.Pipeline. Otherwise it has no effect.

Parameters: data (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for data parameter in transform.
Returns: self – The updated object.
Return type: object

transform(data)¶

Given one “original” dataframe, discretise it.

Parameters: data (DataFrame) – dataframe with continuous features, to be transformed into discrete
Return type: array
Returns: discretised version of the input data