actableai.intervention package
Submodules
actableai.intervention.config module
actableai.intervention.model module
-
class
actableai.intervention.model.AAIInterventionEffectPredictor(target: str, current_intervention_column: str, common_causes: Optional[List[str]] = None, new_intervention_column: Optional[str] = None, expected_target: Optional[str] = None, causal_cv: Optional[int] = None, causal_hyperparameters: Optional[Dict] = None, cate_alpha: Optional[float] = None, presets: Optional[str] = None, model_directory: Optional[str] = None, num_gpus: Optional[int] = 0, drop_unique: bool = True, drop_useless_features: bool = False, tabpfn_model_directory: Optional[str] = None, cross_validation_hyperparameters: Optional[Dict] = None)
Bases: object
-
fit(df: pandas.core.frame.DataFrame) → actableai.intervention.model.AAIInterventionEffectPredictor
- Generate each appropriate models (treatment, outcome and residuals)
- then fit the final DML causal model
- Parameters
- df – DataFrame containing the values to fit on
- Returns
- Self fitted predictor
- Return type
- AAIInterventionEffectPredictor
-
predict(df: pandas.core.frame.DataFrame) → pandas.core.frame.DataFrame
Predict the effect of the treatment on the outcome
- Parameters
- df – DataFrame containing the values to predict on
- Raises
- NotFittedError – If this method is called before fit
- Returns
- DataFrame containing the effect of the treatment on the
- outcome
- Return type
- pd.DataFrame
-
predict_two_way(df: pandas.core.frame.DataFrame) → pandas.core.frame.DataFrame
- Predict effect of the new treatment on the outcome AND predict the treatment
- necessary to obtain an expected outcome
- Parameters
- df – Input DataFrame.
NB : Here df can also contain the “expected_outcome”. When the new
treatment is set the expected outcome must be nan and when the expected
outcome is set the new treatment must be nan
- Raises
- NotFittedError – If this method is called before fit
- Returns
- Result containing the expected outcome when the treatment
- is not nan and the new treatment when the expected outcome is not nan
- Return type
- pd.DataFrame
Module contents