causallift.nodes package

Submodules

causallift.nodes.estimate_propensity module

causallift.nodes.estimate_propensity.estimate_propensity(args, df, model)

causallift.nodes.estimate_propensity.fit_propensity(args, df)

causallift.nodes.estimate_propensity.schedule_propensity_scoring(args, df)

causallift.nodes.model_for_each module

class causallift.nodes.model_for_each.ModelForTreated(*posargs, **kwargs)

Bases: ModelForTreatedOrUntreated

__init__(*posargs, **kwargs)

class causallift.nodes.model_for_each.ModelForTreatedOrUntreated(treatment_val=1.0)

Bases: object

__init__(treatment_val=1.0)

fit(args, df_)

predict_proba(args, df_, models_dict)

simulate_recommendation(args, df_, models_dict)

class causallift.nodes.model_for_each.ModelForUntreated(*posargs, **kwargs)

Bases: ModelForTreatedOrUntreated

__init__(*posargs, **kwargs)

causallift.nodes.model_for_each.bundle_treated_and_untreated_models(treated_model, untreated_model)

causallift.nodes.model_for_each.model_for_treated_fit(*posargs, **kwargs)

causallift.nodes.model_for_each.model_for_treated_predict_proba(*posargs, **kwargs)

causallift.nodes.model_for_each.model_for_treated_simulate_recommendation(*posargs, **kwargs)

causallift.nodes.model_for_each.model_for_untreated_fit(*posargs, **kwargs)

causallift.nodes.model_for_each.model_for_untreated_predict_proba(*posargs, **kwargs)

causallift.nodes.model_for_each.model_for_untreated_simulate_recommendation(*posargs, **kwargs)

causallift.nodes.utils module

causallift.nodes.utils.add_cate_to_df(args, df, cate_estimated, proba_treated, proba_untreated)

causallift.nodes.utils.apply_method(obj, method, **kwargs)

causallift.nodes.utils.bundle_train_and_test_data(args, train_df, test_df)

causallift.nodes.utils.compute_cate(proba_treated, proba_untreated)

causallift.nodes.utils.concat_train_test(args, train, test)

Concatenate train and test series. Use series.xs(‘train’) or series.xs(‘test’) to split

causallift.nodes.utils.concat_train_test_df(args, train, test)

Concatenate train and test data frames. Use df.xs(‘train’) or df.xs(‘test’) to split.

causallift.nodes.utils.conf_mat_df(y_true, y_pred)

causallift.nodes.utils.estimate_effect(args, sim_treated_df, sim_untreated_df)

causallift.nodes.utils.gain_tuple(df_, r_)

causallift.nodes.utils.get_cols_features(df, non_feature_cols=['Treatment', 'Outcome', 'TransformedOutcome', 'Propensity', 'Recommendation'])

causallift.nodes.utils.impute_cols_features(args, df)

causallift.nodes.utils.initialize_model(args, model_key='uplift_model_params', default_estimator='sklearn.linear_model.LogisticRegression')

Return type:

Type[BaseEstimator]

causallift.nodes.utils.len_o(df, outcome=1.0, col_outcome='Outcome')

causallift.nodes.utils.len_t(df, treatment=1.0, col_treatment='Treatment')

causallift.nodes.utils.len_to(df, treatment=1.0, outcome=1.0, col_treatment='Treatment', col_outcome='Outcome')

causallift.nodes.utils.outcome_fraction_(df, col_outcome='Outcome')

causallift.nodes.utils.overall_uplift_gain_(df, treatment=1.0, outcome=1.0, col_treatment='Treatment', col_outcome='Outcome')

causallift.nodes.utils.recommend_by_cate(args, df, treatment_fractions)

causallift.nodes.utils.score_df(y_train, y_test, y_pred_train, y_pred_test, average='binary')

causallift.nodes.utils.treatment_fraction_(df, col_treatment='Treatment')

causallift.nodes.utils.treatment_fractions_(args, df)

Return type:

Type[EasyDict]