can now optionally specify treatment and control values for estimate effect--useful for continuous tratments

dowhy/causal_estimator.py

@@ -10,7 +10,8 @@ class CausalEstimator:
     """
 
     def __init__(self, data, identified_estimand, treatment, outcome,
-                 test_significance, evaluate_effect_strength=False,
+                 control_value=0, treatment_value=1,
+                 test_significance=False, evaluate_effect_strength=False,
                  confidence_intervals = False,
                  target_units=None, effect_modifiers=None,
                  params=None):
@@ -36,6 +37,8 @@ class CausalEstimator:
         # Currently estimation methods only support univariate treatment and outcome
         self._treatment_name = treatment[0]
         self._outcome_name = outcome[0]
+        self._control_value = control_value
+        self._treatment_value = treatment_value
         self._significance_test = test_significance
         self._effect_strength_eval = evaluate_effect_strength
         self._target_units = target_units

dowhy/causal_estimators/econml_cate_estimator.py

@@ -76,8 +76,8 @@ class EconmlCateEstimator(CausalEstimator):
             boolean_criterion = np.array(filtered_rows.notnull().iloc[:,0])
             X_test = X[boolean_criterion]
             n_target_units = X_test.shape[0]
-        T0_test = np.zeros((n_target_units, 1)) # single-dimensional treatment
-        T1_test = np.ones((n_target_units, 1))
+        T0_test = np.repeat([[self._control_value]], n_target_units, axis=0)
+        T1_test = np.repeat([[self._treatment_value]], n_target_units, axis=0)
         est = self.estimator.effect(X_test, T0 = T0_test, T1 = T1_test)
         ate = np.mean(est)
 

dowhy/causal_estimators/linear_regression_estimator.py

@@ -35,7 +35,7 @@ class LinearRegressionEstimator(CausalEstimator):
         coefficients = self._linear_model.coef_
         self.logger.debug("Coefficients of the fitted linear model: " +
                           ",".join(map(str, coefficients)))
-        effect_estimate = self._do(1) - self._do(0)
+        effect_estimate = self._do(self._treatment_value) - self._do(self._control_value)
         estimate = CausalEstimate(estimate=effect_estimate,
                               target_estimand=self._target_estimand,
                               realized_estimand_expr=self.symbolic_estimator,

dowhy/causal_model.py

@@ -135,6 +135,8 @@ class CausalModel:
         return identified_estimand
 
     def estimate_effect(self, identified_estimand, method_name=None,
+                        control_value = 0,
+                        treatment_value = 1,
                         test_significance=None, evaluate_effect_strength=False,
                         confidence_intervals=False,
                         target_units="ate", effect_modifiers=None,
@@ -180,6 +182,8 @@ class CausalModel:
                 self._data,
                 identified_estimand,
                 self._treatment, self._outcome, #names of treatment and outcome
+                control_value = control_value,
+                treatment_value = treatment_value,
                 test_significance=test_significance,
                 evaluate_effect_strength=evaluate_effect_strength,
                 confidence_intervals = confidence_intervals,