def test_inverse_transform(self):
# We need a lot of components for the reconstruction to be "almost
# equal" in all positions. XXX Test means or sums instead?
tsvd = TruncatedSVD(n_components=52,
random_state=42,
algorithm='randomized')
Xt = tsvd.fit_transform(self.X)
Xinv = tsvd.inverse_transform(Xt)
assert_array_almost_equal(Xinv.fetch(), self.Xdense, decimal=1)
def test_singular_values(self):
# Check that the TruncatedSVD output has the correct singular values
# Set the singular values and see what we get back
rng = np.random.RandomState(0)
n_samples = 100
n_features = 110
X = rng.randn(n_samples, n_features)
rpca = TruncatedSVD(n_components=3,
algorithm='randomized',
random_state=rng)
X_rpca = rpca.fit_transform(X)
X_rpca /= mt.sqrt(mt.sum(X_rpca**2.0, axis=0))
X_rpca[:, 0] *= 3.142
X_rpca[:, 1] *= 2.718
X_hat_rpca = mt.dot(X_rpca, rpca.components_)
rpca.fit(X_hat_rpca)
assert_array_almost_equal(rpca.singular_values_.execute(),
[3.142, 2.718, 1.0], 14)
def test_explained_variance(self):
# Test sparse data
svd_r_10_sp = TruncatedSVD(10, algorithm="randomized", random_state=42)
svd_r_20_sp = TruncatedSVD(20, algorithm="randomized", random_state=42)
X_trans_r_10_sp = svd_r_10_sp.fit_transform(self.X)
X_trans_r_20_sp = svd_r_20_sp.fit_transform(self.X)
# Test dense data
svd_r_10_de = TruncatedSVD(10, algorithm="randomized", random_state=42)
svd_r_20_de = TruncatedSVD(20, algorithm="randomized", random_state=42)
X_trans_r_10_de = svd_r_10_de.fit_transform(self.X.toarray())
X_trans_r_20_de = svd_r_20_de.fit_transform(self.X.toarray())
# helper arrays for tests below
svds = (svd_r_10_sp, svd_r_20_sp, svd_r_10_de, svd_r_20_de)
svds_trans = (
(svd_r_10_sp, X_trans_r_10_sp),
(svd_r_20_sp, X_trans_r_20_sp),
(svd_r_10_de, X_trans_r_10_de),
(svd_r_20_de, X_trans_r_20_de),
)
svds_10_v_20 = (
(svd_r_10_sp, svd_r_20_sp),
(svd_r_10_de, svd_r_20_de),
)
svds_sparse_v_dense = (
(svd_r_10_sp, svd_r_10_de),
(svd_r_20_sp, svd_r_20_de),
)
# Assert the 1st component is equal
for svd_10, svd_20 in svds_10_v_20:
assert_array_almost_equal(
svd_10.explained_variance_ratio_.execute(),
svd_20.explained_variance_ratio_[:10].execute(),
decimal=4,
)
# Assert that 20 components has higher explained variance than 10
for svd_10, svd_20 in svds_10_v_20:
self.assertGreater(
svd_20.explained_variance_ratio_.sum().execute(),
svd_10.explained_variance_ratio_.sum().execute(),
)
# Assert that all the values are greater than 0
for svd in svds:
assert_array_less(0.0, svd.explained_variance_ratio_.execute())
# Assert that total explained variance is less than 1
for svd in svds:
assert_array_less(svd.explained_variance_ratio_.sum().execute(),
1.0)
# Compare sparse vs. dense
for svd_sparse, svd_dense in svds_sparse_v_dense:
assert_array_almost_equal(
svd_sparse.explained_variance_ratio_.execute(),
svd_dense.explained_variance_ratio_.execute())
# Test that explained_variance is correct
for svd, transformed in svds_trans:
total_variance = mt.var(self.X.toarray(), axis=0).sum().execute()
variances = mt.var(transformed, axis=0)
true_explained_variance_ratio = variances / total_variance
assert_array_almost_equal(
svd.explained_variance_ratio_.execute(),
true_explained_variance_ratio.execute(),
)
def test_integers(self):
Xint = self.X.astype(np.int64)
tsvd = TruncatedSVD(n_components=6)
Xtrans = tsvd.fit_transform(Xint)
self.assertEqual(Xtrans.shape, (self.n_samples, tsvd.n_components))
def test_sparse_formats(self):
tsvd = TruncatedSVD(n_components=11)
Xtrans = tsvd.fit_transform(self.Xdense)
self.assertEqual(Xtrans.shape, (self.n_samples, 11))
Xtrans = tsvd.transform(self.Xdense)
self.assertEqual(Xtrans.shape, (self.n_samples, 11))
def test_too_many_components(self):
for n_components in (self.n_features, self.n_features + 1):
tsvd = TruncatedSVD(n_components=n_components,
algorithm='randomized')
with self.assertRaises(ValueError):
tsvd.fit(self.X)
def test_attributes(self):
for n_components in (10, 25, 41):
tsvd = TruncatedSVD(n_components).fit(self.X)
self.assertEqual(tsvd.n_components, n_components)
self.assertEqual(tsvd.components_.shape,
(n_components, self.n_features))
