def test_drlearner(self):
y, T, X, W = self._get_data()
for est in [
LinearDRLearner(random_state=123),
SparseLinearDRLearner(random_state=123)
]:
est.fit(y, T, X=X, W=W, cache_values=True)
np.testing.assert_equal(est.model_regression, 'auto')
est.model_regression = LinearRegression()
est.model_propensity = LogisticRegression(random_state=123)
est.fit(y, T, X=X, W=W, cache_values=True)
assert isinstance(est.model_regression, LinearRegression)
with pytest.raises(ValueError):
est.multitask_model_final = True
with pytest.raises(ValueError):
est.model_final = LinearRegression()
est.min_propensity = .1
est.mc_iters = 2
est.featurizer = PolynomialFeatures(degree=2, include_bias=False)
est.refit_final()
assert isinstance(est.featurizer_, PolynomialFeatures)
np.testing.assert_equal(est.mc_iters, 2)
intcpt = est.intercept_(T=1)
est.fit_cate_intercept = False
np.testing.assert_equal(est.intercept_(T=1), intcpt)
est.refit_final()
with pytest.raises(AttributeError):
est.intercept(T=1)
est.fit(y, T, X=X, W=W, cache_values=False)
with pytest.raises(AssertionError):
est.refit_final()
def test_can_summarize(self):
LinearDML(model_t=LinearRegression(),
model_y=LinearRegression()).fit(TestInference.Y,
TestInference.T,
TestInference.X,
TestInference.W).summary()
LinearDRLearner(model_regression=LinearRegression(),
model_propensity=LogisticRegression(),
fit_cate_intercept=False).fit(
TestInference.Y,
TestInference.T > 0,
TestInference.X,
TestInference.W,
inference=BootstrapInference(5)).summary(1)
def test_store_dataframe_name(self):
Y, T, X, W, Z = self._get_data()
Y_name = "outcome"
Y = pd.Series(Y, name=Y_name)
T_name = "treatment"
T = pd.Series(T, name=T_name)
X_name = ["feature"]
X = pd.DataFrame(X, columns=X_name)
W_name = ["control1", "control2", "control3", "control4"]
W = pd.DataFrame(W, columns=W_name)
est = LinearDRLearner().dowhy.fit(Y, T, X, W)
np.testing.assert_array_equal(est._common_causes, X_name + W_name)
np.testing.assert_array_equal(est._effect_modifiers, X_name)
np.testing.assert_array_equal(est._treatment, [T_name])
np.testing.assert_array_equal(est._outcome, [Y_name])
def test_dr_random_state(self):
Y, T, X, W, X_test = self._make_data(500, 2)
for est in [
DRLearner(model_final=RandomForestRegressor(
max_depth=3,
n_estimators=10,
min_samples_leaf=100,
bootstrap=True,
random_state=123),
cv=2,
random_state=123),
LinearDRLearner(random_state=123),
SparseLinearDRLearner(cv=2, random_state=123),
ForestDRLearner(
model_regression=RandomForestRegressor(n_estimators=10,
max_depth=4,
random_state=123),
model_propensity=RandomForestClassifier(n_estimators=10,
max_depth=4,
random_state=123),
cv=2,
random_state=123)
]:
TestRandomState._test_random_state(est, X_test, Y, T, X=X, W=W)
def test_mean_pred_stderr(self):
"""Test that mean_pred_stderr is not None when estimator's final stage is linear"""
Y, T, X, W = TestInference.Y, TestInference.T, TestInference.X, TestInference.W
ests = [
LinearDML(model_t=LinearRegression(),
model_y=LinearRegression(),
featurizer=PolynomialFeatures(degree=2,
include_bias=False)),
LinearDRLearner(model_regression=LinearRegression(),
model_propensity=LogisticRegression(),
featurizer=PolynomialFeatures(degree=2,
include_bias=False))
]
for est in ests:
est.fit(Y, T, X=X, W=W)
assert est.const_marginal_effect_inference(
X).population_summary().mean_pred_stderr is not None
# only is not None when T1 is a constant or a list of constant
assert est.effect_inference(
X).population_summary().mean_pred_stderr is not None
if est.__class__.__name__ == "LinearDRLearner":
assert est.coef__inference(T=1).mean_pred_stderr is None
else:
assert est.coef__inference().mean_pred_stderr is None
def test_summary_discrete(self):
"""Tests the inference results summary for discrete treatment estimators."""
# Test inference results when `cate_feature_names` doesn not exist
for inference in [BootstrapInference(n_bootstrap_samples=5), 'auto']:
cate_est = LinearDRLearner(model_regression=LinearRegression(), model_propensity=LogisticRegression(),
featurizer=PolynomialFeatures(degree=2,
include_bias=False)
)
cate_est.fit(
TestInference.Y,
TestInference.T,
TestInference.X,
TestInference.W,
inference=inference
)
summary_results = cate_est.summary(T=1)
coef_rows = np.asarray(summary_results.tables[0].data)[1:, 0]
default_names = get_input_columns(TestInference.X)
fnames = PolynomialFeatures(degree=2, include_bias=False).fit(
TestInference.X).get_feature_names(default_names)
np.testing.assert_array_equal(coef_rows, fnames)
intercept_rows = np.asarray(summary_results.tables[1].data)[1:, 0]
np.testing.assert_array_equal(intercept_rows, ['cate_intercept'])
cate_est = LinearDRLearner(model_regression=LinearRegression(),
model_propensity=LogisticRegression(),
featurizer=PolynomialFeatures(degree=2,
include_bias=False)
)
cate_est.fit(
TestInference.Y,
TestInference.T,
TestInference.X,
TestInference.W,
inference=inference
)
fnames = ['Q' + str(i) for i in range(TestInference.d_x)]
summary_results = cate_est.summary(T=1, feature_names=fnames)
coef_rows = np.asarray(summary_results.tables[0].data)[1:, 0]
fnames = PolynomialFeatures(degree=2, include_bias=False).fit(
TestInference.X).get_feature_names(input_features=fnames)
np.testing.assert_array_equal(coef_rows, fnames)
cate_est = LinearDRLearner(model_regression=LinearRegression(),
model_propensity=LogisticRegression(), featurizer=None)
cate_est.fit(
TestInference.Y,
TestInference.T,
TestInference.X,
TestInference.W,
inference=inference
)
summary_results = cate_est.summary(T=1)
coef_rows = np.asarray(summary_results.tables[0].data)[1:, 0]
np.testing.assert_array_equal(coef_rows, ['X' + str(i) for i in range(TestInference.d_x)])
cate_est = LinearDRLearner(model_regression=LinearRegression(),
model_propensity=LogisticRegression(), featurizer=None)
cate_est.fit(
TestInference.Y,
TestInference.T,
TestInference.X,
TestInference.W,
inference=inference
)
fnames = ['Q' + str(i) for i in range(TestInference.d_x)]
summary_results = cate_est.summary(T=1, feature_names=fnames)
coef_rows = np.asarray(summary_results.tables[0].data)[1:, 0]
np.testing.assert_array_equal(coef_rows, fnames)
cate_est = LinearDRLearner(model_regression=LinearRegression(),
model_propensity=LogisticRegression(), featurizer=None)
wrapped_est = self._NoFeatNamesEst(cate_est)
wrapped_est.fit(
TestInference.Y,
TestInference.T,
TestInference.X,
TestInference.W,
inference=inference
)
summary_results = wrapped_est.summary(T=1)
coef_rows = np.asarray(summary_results.tables[0].data)[1:, 0]
np.testing.assert_array_equal(coef_rows, ['X' + str(i) for i in range(TestInference.d_x)])
cate_est = LinearDRLearner(model_regression=LinearRegression(),
model_propensity=LogisticRegression(), featurizer=None)
wrapped_est = self._NoFeatNamesEst(cate_est)
wrapped_est.fit(
TestInference.Y,
TestInference.T,
TestInference.X,
TestInference.W,
inference=inference
)
fnames = ['Q' + str(i) for i in range(TestInference.d_x)]
summary_results = wrapped_est.summary(T=1, feature_names=fnames)
coef_rows = np.asarray(summary_results.tables[0].data)[1:, 0]
np.testing.assert_array_equal(coef_rows, fnames)