dice_ml package

Subpackages

Submodules

dice_ml.constants module

Constants for dice-ml package.

class dice_ml.constants.BackEndTypes[source]

Bases: object

ALL = ['sklearn', 'TF1', 'TF2', 'PYT']

Pytorch = 'PYT'

Sklearn = 'sklearn'

Tensorflow1 = 'TF1'

Tensorflow2 = 'TF2'

class dice_ml.constants.ModelTypes[source]

Bases: object

ALL = ['classifier', 'regressor']

Classifier = 'classifier'

Regressor = 'regressor'

class dice_ml.constants.SamplingStrategy[source]

Bases: object

Genetic = 'genetic'

Gradient = 'gradient'

KdTree = 'kdtree'

Random = 'random'

dice_ml.counterfactual_explanations module

class dice_ml.counterfactual_explanations.CounterfactualExplanations(cf_examples_list, local_importance=None, summary_importance=None, version=None)[source]

Bases: object

A class to store counterfactual examples for one or more inputs and feature importance scores.

Parameters

cf_examples_list – A list of CounterfactualExamples instances
local_importance – List of estimated local importance scores. The size of the list is the number of input instances, each containing feature importance scores for that input.
summary_importance – Estimated global feature importance scores based on the input set of CounterfactualExamples instances

property cf_examples_list

static from_json(json_str)[source]: Deserialize json string to a CounterfactualExplanations object.

property local_importance

property metadata

property summary_importance

to_json()[source]: Serialize Explanations object to json.

visualize_as_dataframe(display_sparse_df=True, show_only_changes=False)[source]

visualize_as_list(display_sparse_df=True, show_only_changes=False)[source]

dice_ml.data module

Module pointing to different implementations of Data class

DiCE requires only few parameters about the data such as the range of continuous features and the levels of categorical features. Hence, DiCE can be used for a private data whose meta data are only available (such as the feature names and range/levels of different features) by specifying appropriate parameters.

class dice_ml.data.Data(**params)[source]

Bases: _BaseData

Class containing all required information about the data for DiCE.

decide_implementation_type(params)[source]: Decides if the Data class is for public or private data.

dice_ml.data.decide(params)[source]

Decides if the Data class is for public or private data.

To add new implementations of Data, add the class in data_interfaces subpackage and import-and-return the class in an elif loop as shown in the below method.

dice_ml.dice module

Module pointing to different implementations of DiCE based on different frameworks such as Tensorflow or PyTorch or sklearn, and different methods such as RandomSampling, DiCEKD or DiCEGenetic

class dice_ml.dice.Dice(data_interface, model_interface, method='random', **kwargs)[source]

Bases: ExplainerBase

An interface class to different DiCE implementations.

decide_implementation_type(data_interface, model_interface, method, **kwargs)[source]: Decides DiCE implementation type.

dice_ml.dice.decide(model_interface, method)[source]

Decides DiCE implementation type.

To add new implementations of DiCE, add the class in explainer_interfaces subpackage and import-and-return the class in an elif loop as shown in the below method.

dice_ml.diverse_counterfactuals module

class dice_ml.diverse_counterfactuals.CounterfactualExamples(data_interface=None, final_cfs_df=None, test_instance_df=None, final_cfs_df_sparse=None, posthoc_sparsity_param=0, desired_range=None, desired_class='opposite', model_type='classifier')[source]

Bases: object

A class to store and visualize the resulting counterfactual explanations.

display_df(df, show_only_changes)[source]

static from_json(cf_example_json_str)[source]

print_list(li, show_only_changes)[source]

to_json(serialization_version)[source]

visualize_as_dataframe(display_sparse_df=True, show_only_changes=False)[source]

visualize_as_list(display_sparse_df=True, show_only_changes=False)[source]

dice_ml.diverse_counterfactuals.json_converter(obj)[source]: Helper function to convert object to json.

dice_ml.model module

Module pointing to different implementations of Model class

The implementations contain methods to access the output or gradients of ML models trained based on different frameworks such as Tensorflow or PyTorch.

class dice_ml.model.Model(model=None, model_path='', backend='TF1', model_type='classifier', func=None, kw_args=None)[source]

Bases: object

An interface class to different ML Model implementations.

decide_implementation_type(model, model_path, backend, func, kw_args)[source]: Decides the Model implementation type.

dice_ml.model.decide(backend)[source]

Decides the Model implementation type.

To add new implementations of Model, add the class in model_interfaces subpackage and import-and-return the class in an elif loop as shown in the below method.

Module contents

class dice_ml.Data(**params)[source]

Bases: _BaseData

Class containing all required information about the data for DiCE.

decide_implementation_type(params)[source]: Decides if the Data class is for public or private data.

class dice_ml.Dice(data_interface, model_interface, method='random', **kwargs)[source]

Bases: ExplainerBase

An interface class to different DiCE implementations.

decide_implementation_type(data_interface, model_interface, method, **kwargs)[source]: Decides DiCE implementation type.

class dice_ml.Model(model=None, model_path='', backend='TF1', model_type='classifier', func=None, kw_args=None)[source]

Bases: object

An interface class to different ML Model implementations.

decide_implementation_type(model, model_path, backend, func, kw_args)[source]: Decides the Model implementation type.