def test_logistic_loss_derivative(n_samples=4, n_features=10, decimal=5):
rng = np.random.RandomState(42)
X = rng.randn(n_samples, n_features)
y = rng.randn(n_samples)
n_features = X.shape[1]
w = rng.randn(n_features + 1)
np.testing.assert_almost_equal(check_grad(
lambda w: _logistic(X, y, w),
lambda w: _logistic_loss_grad(X, y, w), w), 0., decimal=decimal)
np.testing.assert_almost_equal(check_grad(
lambda w: _logistic(X, y, w),
lambda w: _logistic_loss_grad(X, y, w), w), 0., decimal=decimal)
def test_logistic_loss_derivative(n_samples=4, n_features=10, decimal=5):
rng = np.random.RandomState(42)
X = rng.randn(n_samples, n_features)
y = rng.randn(n_samples)
n_features = X.shape[1]
w = rng.randn(n_features + 1)
np.testing.assert_almost_equal(
check_grad(lambda w: _logistic(X, y, w), lambda w: _logistic_loss_grad(X, y, w), w), 0.0, decimal=decimal
)
np.testing.assert_almost_equal(
check_grad(lambda w: _logistic(X, y, w), lambda w: _logistic_loss_grad(X, y, w), w), 0.0, decimal=decimal
)
def test_logistic_lipschitz(n_samples=4, n_features=2, random_state=42):
rng = np.random.RandomState(random_state)
for scaling in np.logspace(-3, 3, num=7):
X = rng.randn(n_samples, n_features) * scaling
y = rng.randn(n_samples)
n_features = X.shape[1]
L = _logistic_loss_lipschitz_constant(X)
_check_lipschitz_continuous(lambda w: _logistic(X, y, w),
n_features + 1, L)
def test_logistic_lipschitz(n_samples=4, n_features=2, random_state=42):
rng = np.random.RandomState(random_state)
for scaling in np.logspace(-3, 3, num=7):
X = rng.randn(n_samples, n_features) * scaling
y = rng.randn(n_samples)
n_features = X.shape[1]
L = _logistic_loss_lipschitz_constant(X)
_check_lipschitz_continuous(lambda w: _logistic(
X, y, w), n_features + 1, L)
