def test_elnet_n_lambda():
rng = np.random.default_rng(SEED)
error = rng.normal(loc=0, scale=1, size=100)
X = rng.normal(loc=5, scale=2, size=(100, 4))
true_betas = np.array([1, -2, 0.5, 1])
y = X.dot(true_betas) + error
elnet = Elnet()
elnet.fit(X, y)
def test_predict_custom_lambda():
rng = np.random.default_rng(SEED)
error = rng.normal(loc=0, scale=1, size=100)
X = rng.normal(loc=5, scale=2, size=(100, 4))
true_betas = np.array([1, -2, 0.5, 1])
y = X.dot(true_betas) + error
m = Elnet(n_splits=3)
m.fit(X, y)
m.predict(X, lamb=1)
def test_predict_without_cv():
rng = np.random.default_rng(SEED)
error = rng.normal(loc=0, scale=1, size=100)
X = rng.normal(loc=5, scale=2, size=(100, 4))
true_betas = np.array([1, -2, 0.5, 1])
y = X.dot(true_betas) + error
m = Elnet(n_splits=1)
m.fit(X, y)
with pytest.raises(Exception):
m.predict(X)
def test_predict_multiple_lambdas():
rng = np.random.default_rng(SEED)
error = rng.normal(loc=0, scale=1, size=100)
X = rng.normal(loc=5, scale=2, size=(100, 4))
true_betas = np.array([1, -2, 0.5, 1])
y = X.dot(true_betas) + error
m = Elnet(n_splits=3)
m.fit(X, y)
preds = m.predict(X, lamb=[0.5, 1, 1.5])
assert preds.shape == (100, 3)
def test_fit_cv_glmnet_comparison():
rng = np.random.default_rng(SEED)
error = rng.normal(loc=0, scale=1, size=100)
X = rng.normal(loc=5, scale=2, size=(100, 4))
true_betas = np.array([1, -2, 0.5, 1])
y = X.dot(true_betas) + error
m = Elnet(n_splits=3, random_state=182, scoring="r2")
m.fit(X, y)
m2 = ElasticNet(n_splits=3, random_state=182)
m2.fit(X, y)
np.testing.assert_almost_equal(m.lambda_max_, m2.lambda_max_)
np.testing.assert_almost_equal(m.lambda_1se_, m2.lambda_best_[0])
def test_fit_cv():
rng = np.random.default_rng(SEED)
error = rng.normal(loc=0, scale=1, size=100)
X = rng.normal(loc=5, scale=2, size=(100, 4))
true_betas = np.array([1, -2, 0.5, 1])
y = X.dot(true_betas) + error
m = Elnet(n_splits=3)
m.fit(X, y)
assert hasattr(m, "lambda_1se_")
assert hasattr(m, "lambda_max_")
assert hasattr(m, "cv_mean")
assert hasattr(m, "cv_se")
def test_elnet_glmnet(alpha):
# compare to original glmnet Fortran algorithm
rng = np.random.default_rng(SEED)
error = rng.normal(loc=0, scale=1, size=100)
X = rng.normal(loc=5, scale=2, size=(100, 4))
true_betas = np.array([1, -2, 0.5, 1])
y = X.dot(true_betas) + error
# glmnet
m = ElasticNet(alpha=alpha)
m.fit(X, y)
# own implementation
m2 = Elnet(alpha=alpha)
m2.fit(X, y)
# same lambda sequence
np.testing.assert_almost_equal(m.lambda_path_, m2.lambda_path_)
# same feature coefficients
# decimal = 4: almost equal to the 4th decimal
np.testing.assert_almost_equal(m.coef_path_, m2.coef_path_, decimal=4)
# same intercept path
np.testing.assert_almost_equal(m.intercept_path_, m2.intercept_path_, decimal=4)