def test_can_use_vectors(self):
"""
TODO Almost identical to DML test, so consider merging
Test that we can pass vectors for T and Y (not only 2-dimensional arrays).
"""
dml = LinearDRLearner(model_regression=LinearRegression(),
model_propensity=LogisticRegression(C=1000, solver='lbfgs', multi_class='auto'),
fit_cate_intercept=False,
featurizer=FunctionTransformer(validate=True))
dml.fit(np.array([1, 2, 1, 2]), np.array([1, 2, 1, 2]), X=np.ones((4, 1)))
self.assertAlmostEqual(dml.coef_(T=2).reshape(())[()], 1)
def test_can_use_statsmodel_inference(self):
"""
TODO Almost identical to DML test, so consider merging
Test that we can use statsmodels to generate confidence intervals
"""
dml = LinearDRLearner(model_regression=LinearRegression(),
model_propensity=LogisticRegression(
C=1000, solver='lbfgs', multi_class='auto'))
dml.fit(np.array([2, 3, 1, 3, 2, 1, 1, 1]),
np.array([3, 2, 1, 2, 3, 1, 1, 1]),
np.ones((8, 1)),
inference='statsmodels')
interval = dml.effect_interval(np.ones((9, 1)),
T0=np.array([1, 1, 1, 1, 1, 1, 1, 1,
1]),
T1=np.array([2, 2, 3, 2, 2, 3, 2, 2,
3]),
alpha=0.05)
point = dml.effect(np.ones((9, 1)),
T0=np.array([1, 1, 1, 1, 1, 1, 1, 1, 1]),
T1=np.array([2, 2, 3, 2, 2, 3, 2, 2, 3]))
assert len(interval) == 2
lo, hi = interval
assert lo.shape == hi.shape == point.shape
assert (lo <= point).all()
assert (point <= hi).all()
assert (lo < hi).any(
) # for at least some of the examples, the CI should have nonzero width
interval = dml.const_marginal_effect_interval(np.ones((9, 1)),
alpha=0.05)
point = dml.const_marginal_effect(np.ones((9, 1)))
assert len(interval) == 2
lo, hi = interval
assert lo.shape == hi.shape == point.shape
assert (lo <= point).all()
assert (point <= hi).all()
assert (lo < hi).any(
) # for at least some of the examples, the CI should have nonzero width
interval = dml.coef__interval(T=2, alpha=0.05)
point = dml.coef_(T=2)
assert len(interval) == 2
lo, hi = interval
assert lo.shape == hi.shape == point.shape
assert (lo <= point).all()
assert (point <= hi).all()
assert (lo < hi).any(
) # for at least some of the examples, the CI should have nonzero width
def test_linear_drlearner_all_attributes(self):
from sklearn.ensemble import GradientBoostingClassifier, GradientBoostingRegressor, RandomForestRegressor
from sklearn.linear_model import LinearRegression, LogisticRegression
from econml.utilities import StatsModelsLinearRegression
import scipy.special
np.random.seed(123)
controls = np.random.uniform(-1, 1, size=(5000, 3))
T = np.random.binomial(2, scipy.special.expit(controls[:, 0]))
sigma = 0.01
y = (1 + .5 * controls[:, 0]) * T + controls[:, 0] + np.random.normal(
0, sigma, size=(5000, ))
for X in [None, controls]:
for W in [None, controls]:
for sample_weight, sample_var in [(None, None),
(np.ones(T.shape[0]),
np.zeros(T.shape[0]))]:
for featurizer in [
None,
PolynomialFeatures(degree=2, include_bias=False)
]:
for models in [(GradientBoostingClassifier(),
GradientBoostingRegressor()),
(LogisticRegression(solver='lbfgs',
multi_class='auto'),
LinearRegression())]:
for inference in [
'statsmodels',
StatsModelsInferenceDiscrete(
cov_type='nonrobust')
]:
with self.subTest(X=X,
W=W,
sample_weight=sample_weight,
sample_var=sample_var,
featurizer=featurizer,
models=models,
inference=inference):
est = LinearDRLearner(
model_propensity=models[0],
model_regression=models[1],
featurizer=featurizer)
if (X is None) and (W is None):
with pytest.raises(
AttributeError) as e_info:
est.fit(
y,
T,
X=X,
W=W,
sample_weight=sample_weight,
sample_var=sample_var)
continue
est.fit(y,
T,
X=X,
W=W,
sample_weight=sample_weight,
sample_var=sample_var,
inference=inference)
if X is not None:
lower, upper = est.effect_interval(
X[:3], T0=0, T1=1)
point = est.effect(X[:3], T0=0, T1=1)
truth = 1 + .5 * X[:3, 0]
TestDRLearner._check_with_interval(
truth, point, lower, upper)
lower, upper = est.const_marginal_effect_interval(
X[:3])
point = est.const_marginal_effect(
X[:3])
truth = np.hstack([
1 + .5 * X[:3, [0]],
2 * (1 + .5 * X[:3, [0]])
])
TestDRLearner._check_with_interval(
truth, point, lower, upper)
else:
lower, upper = est.effect_interval(
T0=0, T1=1)
point = est.effect(T0=0, T1=1)
truth = np.array([1])
TestDRLearner._check_with_interval(
truth, point, lower, upper)
lower, upper = est.const_marginal_effect_interval(
)
point = est.const_marginal_effect()
truth = np.array([[1, 2]])
TestDRLearner._check_with_interval(
truth, point, lower, upper)
for t in [1, 2]:
if X is not None:
lower, upper = est.marginal_effect_interval(
t, X[:3])
point = est.marginal_effect(
t, X[:3])
truth = np.hstack([
1 + .5 * X[:3, [0]],
2 * (1 + .5 * X[:3, [0]])
])
TestDRLearner._check_with_interval(
truth, point, lower, upper)
else:
lower, upper = est.marginal_effect_interval(
t)
point = est.marginal_effect(t)
truth = np.array([[1, 2]])
TestDRLearner._check_with_interval(
truth, point, lower, upper)
assert isinstance(est.score_, float)
assert isinstance(
est.score(y, T, X=X, W=W), float)
if X is not None:
feat_names = ['A', 'B', 'C']
else:
feat_names = []
out_feat_names = feat_names
if X is not None:
if (featurizer is not None):
out_feat_names = featurizer.fit(
X).get_feature_names(
feat_names)
np.testing.assert_array_equal(
est.featurizer.
n_input_features_, 3)
np.testing.assert_array_equal(
est.cate_feature_names(feat_names),
out_feat_names)
if isinstance(models[0],
GradientBoostingClassifier):
np.testing.assert_array_equal(
np.array([
mdl.feature_importances_ for
mdl in est.models_regression
]).shape, [
2, 2 + len(feat_names) +
(W.shape[1]
if W is not None else 0)
])
np.testing.assert_array_equal(
np.array([
mdl.feature_importances_ for
mdl in est.models_propensity
]).shape, [
2,
len(feat_names) +
(W.shape[1]
if W is not None else 0)
])
else:
np.testing.assert_array_equal(
np.array([
mdl.coef_ for mdl in
est.models_regression
]).shape, [
2, 2 + len(feat_names) +
(W.shape[1]
if W is not None else 0)
])
np.testing.assert_array_equal(
np.array([
mdl.coef_ for mdl in
est.models_propensity
]).shape, [
2, 3,
len(feat_names) +
(W.shape[1]
if W is not None else 0)
])
if X is not None:
for t in [1, 2]:
true_coef = np.zeros(
len(out_feat_names))
true_coef[0] = .5 * t
lower, upper = est.model_cate(
T=t).coef__interval()
point = est.model_cate(T=t).coef_
truth = true_coef
TestDRLearner._check_with_interval(
truth, point, lower, upper)
lower, upper = est.coef__interval(
t)
point = est.coef_(t)
truth = true_coef
TestDRLearner._check_with_interval(
truth, point, lower, upper)
for t in [1, 2]:
lower, upper = est.model_cate(
T=t).intercept__interval()
point = est.model_cate(T=t).intercept_
truth = t
TestDRLearner._check_with_interval(
truth, point, lower, upper)
lower, upper = est.intercept__interval(
t)
point = est.intercept_(t)
truth = t
TestDRLearner._check_with_interval(
truth, point, lower, upper)
