from typing import Any, List, Optional, Union
import numpy as np
import pandas as pd
import statsmodels.api as sm
from dowhy.causal_estimator import CausalEstimate, CausalEstimator, IdentifiedEstimand
[docs]class RegressionEstimator(CausalEstimator):
"""Compute effect of treatment using some regression function.
Fits a regression model for estimating the outcome using treatment(s) and
confounders.
Base class for all regression models, inherited by
LinearRegressionEstimator and GeneralizedLinearModelEstimator.
"""
def __init__(
self,
identified_estimand: IdentifiedEstimand,
test_significance: bool = False,
evaluate_effect_strength: bool = False,
confidence_intervals: bool = False,
num_null_simulations: int = CausalEstimator.DEFAULT_NUMBER_OF_SIMULATIONS_STAT_TEST,
num_simulations: int = CausalEstimator.DEFAULT_NUMBER_OF_SIMULATIONS_CI,
sample_size_fraction: int = CausalEstimator.DEFAULT_SAMPLE_SIZE_FRACTION,
confidence_level: float = CausalEstimator.DEFAULT_CONFIDENCE_LEVEL,
need_conditional_estimates: Union[bool, str] = "auto",
num_quantiles_to_discretize_cont_cols: int = CausalEstimator.NUM_QUANTILES_TO_DISCRETIZE_CONT_COLS,
**kwargs,
):
"""
:param identified_estimand: probability expression
representing the target identified estimand to estimate.
:param test_significance: Binary flag or a string indicating whether to test significance and by which method. All estimators support test_significance="bootstrap" that estimates a p-value for the obtained estimate using the bootstrap method. Individual estimators can override this to support custom testing methods. The bootstrap method supports an optional parameter, num_null_simulations. If False, no testing is done. If True, significance of the estimate is tested using the custom method if available, otherwise by bootstrap.
:param evaluate_effect_strength: (Experimental) whether to evaluate the strength of effect
:param confidence_intervals: Binary flag or a string indicating whether the confidence intervals should be computed and which method should be used. All methods support estimation of confidence intervals using the bootstrap method by using the parameter confidence_intervals="bootstrap". The bootstrap method takes in two arguments (num_simulations and sample_size_fraction) that can be optionally specified in the params dictionary. Estimators may also override this to implement their own confidence interval method. If this parameter is False, no confidence intervals are computed. If True, confidence intervals are computed by the estimator's specific method if available, otherwise through bootstrap
:param num_null_simulations: The number of simulations for testing the
statistical significance of the estimator
:param num_simulations: The number of simulations for finding the
confidence interval (and/or standard error) for a estimate
:param sample_size_fraction: The size of the sample for the bootstrap
estimator
:param confidence_level: The confidence level of the confidence
interval estimate
:param need_conditional_estimates: Boolean flag indicating whether
conditional estimates should be computed. Defaults to True if
there are effect modifiers in the graph
:param num_quantiles_to_discretize_cont_cols: The number of quantiles
into which a numeric effect modifier is split, to enable
estimation of conditional treatment effect over it.
:param kwargs: (optional) Additional estimator-specific parameters
"""
super().__init__(
identified_estimand=identified_estimand,
test_significance=test_significance,
evaluate_effect_strength=evaluate_effect_strength,
confidence_intervals=confidence_intervals,
num_null_simulations=num_null_simulations,
num_simulations=num_simulations,
sample_size_fraction=sample_size_fraction,
confidence_level=confidence_level,
need_conditional_estimates=need_conditional_estimates,
num_quantiles_to_discretize_cont_cols=num_quantiles_to_discretize_cont_cols,
**kwargs,
)
self.model = None
[docs] def fit(
self,
data: pd.DataFrame,
treatment_name: str,
outcome_name: str,
effect_modifier_names: Optional[List[str]] = None,
):
"""
Fits the estimator with data for effect estimation
:param data: data frame containing the data
:param treatment: name of the treatment variable
:param outcome: name of the outcome variable
:param effect_modifiers: Variables on which to compute separate
effects, or return a heterogeneous effect function. Not all
methods support this currently.
"""
self._set_data(data, treatment_name, outcome_name)
self._set_effect_modifiers(effect_modifier_names)
self.logger.debug("Back-door variables used:" + ",".join(self._target_estimand.get_backdoor_variables()))
self._observed_common_causes_names = self._target_estimand.get_backdoor_variables()
if len(self._observed_common_causes_names) > 0:
self._observed_common_causes = self._data[self._observed_common_causes_names]
self._observed_common_causes = pd.get_dummies(self._observed_common_causes, drop_first=True)
else:
self._observed_common_causes = None
self.symbolic_estimator = self.construct_symbolic_estimator(self._target_estimand)
self.logger.info(self.symbolic_estimator)
# The model is always built on the entire data
_, self.model = self._build_model()
coefficients = self.model.params[1:] # first coefficient is the intercept
self.logger.debug("Coefficients of the fitted model: " + ",".join(map(str, coefficients)))
self.logger.debug(self.model.summary())
return self
[docs] def estimate_effect(
self,
data: pd.DataFrame = None,
treatment_value: Any = 1,
control_value: Any = 0,
target_units=None,
need_conditional_estimates=None,
**_,
):
if data is None:
data = self._data
self._target_units = target_units
self._treatment_value = treatment_value
self._control_value = control_value
# TODO make treatment_value and control value also as local parameters
# All treatments are set to the same constant value
effect_estimate = self._do(treatment_value, data) - self._do(control_value, data)
conditional_effect_estimates = None
if need_conditional_estimates:
conditional_effect_estimates = self._estimate_conditional_effects(
self._estimate_effect_fn, effect_modifier_names=self._effect_modifier_names
)
intercept_parameter = self.model.params[0]
estimate = CausalEstimate(
estimate=effect_estimate,
control_value=control_value,
treatment_value=treatment_value,
conditional_estimates=conditional_effect_estimates,
target_estimand=self._target_estimand,
realized_estimand_expr=self.symbolic_estimator,
intercept=intercept_parameter,
)
estimate.add_estimator(self)
return estimate
def _estimate_effect_fn(self, data_df):
est = self.estimate_effect(data=data_df, need_conditional_estimates=False)
return est.value
def _build_features(self, treatment_values=None, data_df=None):
# Using all data by default
if data_df is None:
data_df = self._data
treatment_vals = pd.get_dummies(self._treatment, drop_first=True)
observed_common_causes_vals = self._observed_common_causes
effect_modifiers_vals = self._effect_modifiers
else:
treatment_vals = pd.get_dummies(data_df[self._treatment_name], drop_first=True)
if len(self._observed_common_causes_names) > 0:
observed_common_causes_vals = data_df[self._observed_common_causes_names]
observed_common_causes_vals = pd.get_dummies(observed_common_causes_vals, drop_first=True)
if self._effect_modifier_names:
effect_modifiers_vals = data_df[self._effect_modifier_names]
effect_modifiers_vals = pd.get_dummies(effect_modifiers_vals, drop_first=True)
# Fixing treatment value to the specified value, if provided
if treatment_values is not None:
treatment_vals = treatment_values
if type(treatment_vals) is not np.ndarray:
treatment_vals = treatment_vals.to_numpy()
# treatment_vals and data_df should have same number of rows
if treatment_vals.shape[0] != data_df.shape[0]:
raise ValueError("Provided treatment values and dataframe should have the same length.")
# Bulding the feature matrix
n_treatment_cols = 1 if len(treatment_vals.shape) == 1 else treatment_vals.shape[1]
n_samples = treatment_vals.shape[0]
treatment_2d = treatment_vals.reshape((n_samples, n_treatment_cols))
if len(self._observed_common_causes_names) > 0:
features = np.concatenate((treatment_2d, observed_common_causes_vals), axis=1)
else:
features = treatment_2d
if self._effect_modifier_names:
for i in range(treatment_2d.shape[1]):
curr_treatment = treatment_2d[:, i]
new_features = curr_treatment[:, np.newaxis] * effect_modifiers_vals.to_numpy()
features = np.concatenate((features, new_features), axis=1)
features = features.astype(
float, copy=False
) # converting to float in case of binary treatment and no other variables
features = sm.add_constant(features, has_constant="add") # to add an intercept term
return features
def _do(self, treatment_val, data_df=None):
if data_df is None:
data_df = self._data
if not self.model:
# The model is always built on the entire data
_, self.model = self._build_model()
# Replacing treatment values by given x
# First, create interventional tensor in original space
interventional_treatment_values = np.full((data_df.shape[0], len(self._treatment_name)), treatment_val)
# Then, use pandas to ensure that the dummies are assigned correctly for a categorical treatment
interventional_treatment_2d = pd.concat(
[
self._treatment.copy(),
pd.DataFrame(data=interventional_treatment_values, columns=self._treatment.columns),
],
axis=0,
).astype(self._treatment.dtypes, copy=False)
interventional_treatment_2d = pd.get_dummies(interventional_treatment_2d, drop_first=True)
interventional_treatment_2d = interventional_treatment_2d[self._treatment.shape[0] :]
new_features = self._build_features(treatment_values=interventional_treatment_2d, data_df=data_df)
interventional_outcomes = self.predict_fn(self.model, new_features)
return interventional_outcomes.mean()