def test_compare_with_sklearn(svd_solver, batch_number):
X = iris.data
X_da = da.from_array(X, chunks=(3, -1))
batch_size = X.shape[0] // batch_number
ipca = sd.IncrementalPCA(n_components=2, batch_size=batch_size)
ipca.fit(X)
ipca_da = IncrementalPCA(n_components=2,
batch_size=batch_size,
svd_solver=svd_solver)
ipca_da.fit(X_da)
np.testing.assert_allclose(
flip_vector_signs(ipca.components_, 1),
flip_vector_signs(ipca_da.components_, 1),
atol=1e-13,
)
np.testing.assert_allclose(ipca.explained_variance_,
ipca_da.explained_variance_,
atol=1e-13)
np.testing.assert_allclose(ipca.explained_variance_,
ipca_da.explained_variance_,
atol=1e-13)
np.testing.assert_allclose(ipca.explained_variance_ratio_,
ipca_da.explained_variance_ratio_,
atol=1e-13)
if svd_solver == "randomized":
# noise variance in randomized solver is probabilistic.
assert_almost_equal(ipca.noise_variance_,
ipca_da.noise_variance_,
decimal=1)
else:
np.testing.assert_allclose(ipca.noise_variance_,
ipca_da.noise_variance_,
atol=1e-13)
def test_basic():
a = dd.PCA()
b = sd.PCA()
a.fit(dX)
b.fit(X)
assert_estimator_equal(a, b, exclude=["components_"])
np.testing.assert_allclose(
flip_vector_signs(a.components_, 1), flip_vector_signs(b.components_, 1)
)
def test_algorithms():
svd_a = sd.TruncatedSVD(30, algorithm="arpack")
svd_r = dd.TruncatedSVD(30, algorithm="tsqr", random_state=42)
Xa = svd_a.fit_transform(Xdense)[:, :6]
Xr = svd_r.fit_transform(dXdense)[:, :6]
assert_array_almost_equal(flip_vector_signs(Xa, axis=0),
flip_vector_signs(Xr, axis=0),
decimal=5)
comp_a = np.abs(svd_a.components_)
comp_r = np.abs(svd_r.components_)
# All elements are equal, but some elements are more equal than others.
assert_array_almost_equal(comp_a[:9], comp_r[:9])
assert_array_almost_equal(comp_a[9:], comp_r[9:], decimal=2)
def test_basic(algorithm):
a = dd.TruncatedSVD(random_state=0, algorithm=algorithm)
b = sd.TruncatedSVD(random_state=0)
b.fit(Xdense)
a.fit(dXdense)
np.testing.assert_allclose(
flip_vector_signs(a.components_, axis=1),
flip_vector_signs(b.components_, axis=1),
atol=1e-3,
)
assert_estimator_equal(
a, b, exclude=["components_", "explained_variance_"], atol=1e-3
)
assert a.explained_variance_.shape == b.explained_variance_.shape
np.testing.assert_allclose(a.explained_variance_, b.explained_variance_, rtol=0.01)