def test_kernel_diag(kernel):
# Test that diag method of kernel returns consistent results.
K_call_diag = np.diag(kernel(X))
K_diag = kernel.diag(X)
assert_array_almost_equal(K_call_diag, K_diag, 5)
def test_lasso_gives_lstsq_solution():
# Test that Lars Lasso gives least square solution at the end
# of the path
_, _, coef_path_ = linear_model.lars_path(X, y, method='lasso')
coef_lstsq = np.linalg.lstsq(X, y)[0]
assert_array_almost_equal(coef_lstsq, coef_path_[:, -1])
def test_lasso_lars_vs_R_implementation():
# Test that sklearn LassoLars implementation agrees with the LassoLars
# implementation available in R (lars library) under the following
# scenarios:
# 1) fit_intercept=False and normalize=False
# 2) fit_intercept=True and normalize=True
# Let's generate the data used in the bug report 7778
y = np.array(
[-6.45006793, -3.51251449, -8.52445396, 6.12277822, -19.42109366])
x = np.array(
[[0.47299829, 0, 0, 0, 0], [0.08239882, 0.85784863, 0, 0, 0],
[0.30114139, -0.07501577, 0.80895216, 0, 0],
[-0.01460346, -0.1015233, 0.0407278, 0.80338378, 0],
[-0.69363927, 0.06754067, 0.18064514, -0.0803561, 0.40427291]])
X = x.T
###########################################################################
# Scenario 1: Let's compare R vs sklearn when fit_intercept=False and
# normalize=False
###########################################################################
#
# The R result was obtained using the following code:
#
# library(lars)
# model_lasso_lars = lars(X, t(y), type="lasso", intercept=FALSE,
# trace=TRUE, normalize=FALSE)
# r = t(model_lasso_lars$beta)
#
r = np.array([[
0, 0, 0, 0, 0, -79.810362809499026, -83.528788732782829,
-83.777653739190711, -83.784156932888934, -84.033390591756657
], [0, 0, 0, 0, -0.476624256777266, 0, 0, 0, 0, 0.025219751009936],
[
0, -3.577397088285891, -4.702795355871871,
-7.016748621359461, -7.614898471899412,
-0.336938391359179, 0, 0, 0.001213370600853,
0.048162321585148
],
[
0, 0, 0, 2.231558436628169, 2.723267514525966,
2.811549786389614, 2.813766976061531, 2.817462468949557,
2.817368178703816, 2.816221090636795
],
[
0, 0, -1.218422599914637, -3.457726183014808,
-4.021304522060710, -45.827461592423745,
-47.776608869312305, -47.911561610746404,
-47.914845922736234, -48.039562334265717
]])
model_lasso_lars = linear_model.LassoLars(alpha=0,
fit_intercept=False,
normalize=False)
model_lasso_lars.fit(X, y)
skl_betas = model_lasso_lars.coef_path_
assert_array_almost_equal(r, skl_betas, decimal=12)
###########################################################################
###########################################################################
# Scenario 2: Let's compare R vs sklearn when fit_intercept=True and
# normalize=True
#
# Note: When normalize is equal to True, R returns the coefficients in
# their original units, that is, they are rescaled back, whereas sklearn
# does not do that, therefore, we need to do this step before comparing
# their results.
###########################################################################
#
# The R result was obtained using the following code:
#
# library(lars)
# model_lasso_lars2 = lars(X, t(y), type="lasso", intercept=TRUE,
# trace=TRUE, normalize=TRUE)
# r2 = t(model_lasso_lars2$beta)
r2 = np.array(
[[0, 0, 0, 0, 0], [0, 0, 0, 8.371887668009453, 19.463768371044026],
[0, 0, 0, 0, 9.901611055290553],
[
0, 7.495923132833733, 9.245133544334507, 17.389369207545062,
26.971656815643499
], [0, 0, -1.569380717440311, -5.924804108067312,
-7.996385265061972]])
model_lasso_lars2 = linear_model.LassoLars(alpha=0, normalize=True)
model_lasso_lars2.fit(X, y)
skl_betas2 = model_lasso_lars2.coef_path_
# Let's rescale back the coefficients returned by sklearn before comparing
# against the R result (read the note above)
temp = X - np.mean(X, axis=0)
normx = np.sqrt(np.sum(temp**2, axis=0))
skl_betas2 /= normx[:, np.newaxis]
assert_array_almost_equal(r2, skl_betas2, decimal=12)
def test_fastica_simple(add_noise, seed):
# Test the FastICA algorithm on very simple data.
rng = np.random.RandomState(seed)
# scipy.stats uses the global RNG:
n_samples = 1000
# Generate two sources:
s1 = (2 * np.sin(np.linspace(0, 100, n_samples)) > 0) - 1
s2 = stats.t.rvs(1, size=n_samples)
s = np.c_[s1, s2].T
center_and_norm(s)
s1, s2 = s
# Mixing angle
phi = 0.6
mixing = np.array([[np.cos(phi), np.sin(phi)], [np.sin(phi),
-np.cos(phi)]])
m = np.dot(mixing, s)
if add_noise:
m += 0.1 * rng.randn(2, 1000)
center_and_norm(m)
# function as fun arg
def g_test(x):
return x**3, (3 * x**2).mean(axis=-1)
algos = ['parallel', 'deflation']
nls = ['logcosh', 'exp', 'cube', g_test]
whitening = [True, False]
for algo, nl, whiten in itertools.product(algos, nls, whitening):
if whiten:
k_, mixing_, s_ = fastica(m.T,
fun=nl,
algorithm=algo,
random_state=rng)
with pytest.raises(ValueError):
fastica(m.T, fun=np.tanh, algorithm=algo)
else:
pca = PCA(n_components=2, whiten=True, random_state=rng)
X = pca.fit_transform(m.T)
k_, mixing_, s_ = fastica(X,
fun=nl,
algorithm=algo,
whiten=False,
random_state=rng)
with pytest.raises(ValueError):
fastica(X, fun=np.tanh, algorithm=algo)
s_ = s_.T
# Check that the mixing model described in the docstring holds:
if whiten:
assert_almost_equal(s_, np.dot(np.dot(mixing_, k_), m))
center_and_norm(s_)
s1_, s2_ = s_
# Check to see if the sources have been estimated
# in the wrong order
if abs(np.dot(s1_, s2)) > abs(np.dot(s1_, s1)):
s2_, s1_ = s_
s1_ *= np.sign(np.dot(s1_, s1))
s2_ *= np.sign(np.dot(s2_, s2))
# Check that we have estimated the original sources
if not add_noise:
assert_almost_equal(np.dot(s1_, s1) / n_samples, 1, decimal=2)
assert_almost_equal(np.dot(s2_, s2) / n_samples, 1, decimal=2)
else:
assert_almost_equal(np.dot(s1_, s1) / n_samples, 1, decimal=1)
assert_almost_equal(np.dot(s2_, s2) / n_samples, 1, decimal=1)
# Test FastICA class
_, _, sources_fun = fastica(m.T, fun=nl, algorithm=algo, random_state=seed)
ica = FastICA(fun=nl, algorithm=algo, random_state=seed)
sources = ica.fit_transform(m.T)
assert ica.components_.shape == (2, 2)
assert sources.shape == (1000, 2)
assert_array_almost_equal(sources_fun, sources)
assert_array_almost_equal(sources, ica.transform(m.T))
assert ica.mixing_.shape == (2, 2)
for fn in [np.tanh, "exp(-.5(x^2))"]:
ica = FastICA(fun=fn, algorithm=algo)
with pytest.raises(ValueError):
ica.fit(m.T)
with pytest.raises(TypeError):
FastICA(fun=range(10)).fit(m.T)
def test_factor_analysis():
# Test FactorAnalysis ability to recover the data covariance structure
rng = np.random.RandomState(0)
n_samples, n_features, n_components = 20, 5, 3
# Some random settings for the generative model
W = rng.randn(n_components, n_features)
# latent variable of dim 3, 20 of it
h = rng.randn(n_samples, n_components)
# using gamma to model different noise variance
# per component
noise = rng.gamma(1, size=n_features) * rng.randn(n_samples, n_features)
# generate observations
# wlog, mean is 0
X = np.dot(h, W) + noise
with pytest.raises(ValueError):
FactorAnalysis(svd_method='foo')
fa_fail = FactorAnalysis()
fa_fail.svd_method = 'foo'
with pytest.raises(ValueError):
fa_fail.fit(X)
fas = []
for method in ['randomized', 'lapack']:
fa = FactorAnalysis(n_components=n_components, svd_method=method)
fa.fit(X)
fas.append(fa)
X_t = fa.transform(X)
assert X_t.shape == (n_samples, n_components)
assert_almost_equal(fa.loglike_[-1], fa.score_samples(X).sum())
assert_almost_equal(fa.score_samples(X).mean(), fa.score(X))
diff = np.all(np.diff(fa.loglike_))
assert diff > 0., 'Log likelihood dif not increase'
# Sample Covariance
scov = np.cov(X, rowvar=0., bias=1.)
# Model Covariance
mcov = fa.get_covariance()
diff = np.sum(np.abs(scov - mcov)) / W.size
assert diff < 0.1, "Mean absolute difference is %f" % diff
fa = FactorAnalysis(n_components=n_components,
noise_variance_init=np.ones(n_features))
with pytest.raises(ValueError):
fa.fit(X[:, :2])
def f(x, y):
return np.abs(getattr(x, y)) # sign will not be equal
fa1, fa2 = fas
for attr in ['loglike_', 'components_', 'noise_variance_']:
assert_almost_equal(f(fa1, attr), f(fa2, attr))
fa1.max_iter = 1
fa1.verbose = True
with pytest.warns(ConvergenceWarning):
fa1.fit(X)
# Test get_covariance and get_precision with n_components == n_features
# with n_components < n_features and with n_components == 0
for n_components in [0, 2, X.shape[1]]:
fa.n_components = n_components
fa.fit(X)
cov = fa.get_covariance()
precision = fa.get_precision()
assert_array_almost_equal(np.dot(cov, precision), np.eye(X.shape[1]),
12)
# test rotation
n_components = 2
results, projections = {}, {}
for method in (None, "varimax", 'quartimax'):
fa_var = FactorAnalysis(n_components=n_components, rotation=method)
results[method] = fa_var.fit_transform(X)
projections[method] = fa_var.get_covariance()
for rot1, rot2 in combinations([None, 'varimax', 'quartimax'], 2):
assert not np.allclose(results[rot1], results[rot2])
assert np.allclose(projections[rot1], projections[rot2], atol=3)
with pytest.raises(ValueError):
FactorAnalysis(rotation="not_implemented").fit_transform(X)
# test against R's psych::principal with rotate="varimax"
# (i.e., the values below stem from rotating the components in R)
# R's factor analysis returns quite different values; therefore, we only
# test the rotation itself
factors = np.array([[0.89421016, -0.35854928, -0.27770122, 0.03773647],
[-0.45081822, -0.89132754, 0.0932195, -0.01787973],
[0.99500666, -0.02031465, 0.05426497, -0.11539407],
[0.96822861, -0.06299656, 0.24411001, 0.07540887]])
r_solution = np.array([[0.962, 0.052], [-0.141, 0.989], [0.949, -0.300],
[0.937, -0.251]])
rotated = _ortho_rotation(factors[:, :n_components], method='varimax').T
assert_array_almost_equal(np.abs(rotated), np.abs(r_solution), decimal=3)
def test_lasso_toy():
# Test Lasso on a toy example for various values of alpha.
# When validating this against glmnet notice that glmnet divides it
# against nobs.
X = [[-1], [0], [1]]
Y = [-1, 0, 1] # just a straight line
T = [[2], [3], [4]] # test sample
clf = Lasso(alpha=1e-8)
clf.fit(X, Y)
pred = clf.predict(T)
assert_array_almost_equal(clf.coef_, [1])
assert_array_almost_equal(pred, [2, 3, 4])
assert_almost_equal(clf.dual_gap_, 0)
clf = Lasso(alpha=0.1)
clf.fit(X, Y)
pred = clf.predict(T)
assert_array_almost_equal(clf.coef_, [.85])
assert_array_almost_equal(pred, [1.7, 2.55, 3.4])
assert_almost_equal(clf.dual_gap_, 0)
clf = Lasso(alpha=0.5)
clf.fit(X, Y)
pred = clf.predict(T)
assert_array_almost_equal(clf.coef_, [.25])
assert_array_almost_equal(pred, [0.5, 0.75, 1.])
assert_almost_equal(clf.dual_gap_, 0)
clf = Lasso(alpha=1)
clf.fit(X, Y)
pred = clf.predict(T)
assert_array_almost_equal(clf.coef_, [.0])
assert_array_almost_equal(pred, [0, 0, 0])
assert_almost_equal(clf.dual_gap_, 0)
def test_singular_values(svd_solver):
# Check that the IncrementalPCA output has the correct singular values
rng = np.random.RandomState(0)
n_samples = 1000
n_features = 100
X = datasets.make_low_rank_matrix(
n_samples, n_features, tail_strength=0.0, effective_rank=10, random_state=rng
)
X = da.from_array(X, chunks=[200, -1])
pca = PCA(n_components=10, svd_solver=svd_solver, random_state=rng).fit(X)
ipca = IncrementalPCA(n_components=10, batch_size=100, svd_solver=svd_solver).fit(X)
assert_array_almost_equal(pca.singular_values_, ipca.singular_values_, 2)
# Compare to the Frobenius norm
X_pca = pca.transform(X)
X_ipca = ipca.transform(X)
assert_array_almost_equal(
np.sum(pca.singular_values_ ** 2.0), np.linalg.norm(X_pca, "fro") ** 2.0, 12
)
assert_array_almost_equal(
np.sum(ipca.singular_values_ ** 2.0), np.linalg.norm(X_ipca, "fro") ** 2.0, 2
)
# Compare to the 2-norms of the score vectors
assert_array_almost_equal(
pca.singular_values_, np.sqrt(np.sum(X_pca ** 2.0, axis=0)), 12
)
assert_array_almost_equal(
ipca.singular_values_, np.sqrt(np.sum(X_ipca ** 2.0, axis=0)), 2
)
# Set the singular values and see what we get back
rng = np.random.RandomState(0)
n_samples = 100
n_features = 110
X = datasets.make_low_rank_matrix(
n_samples, n_features, tail_strength=0.0, effective_rank=3, random_state=rng
)
X = da.from_array(X, chunks=[4, -1])
pca = PCA(n_components=3, svd_solver=svd_solver, random_state=rng)
ipca = IncrementalPCA(n_components=3, batch_size=100, svd_solver=svd_solver)
X_pca = pca.fit_transform(X)
X_pca /= np.sqrt(np.sum(X_pca ** 2.0, axis=0))
X_pca[:, 0] *= 3.142
X_pca[:, 1] *= 2.718
X_hat = np.dot(X_pca, pca.components_)
pca.fit(X_hat)
X_hat = da.from_array(X_hat, chunks=(4, -1))
ipca.fit(X_hat)
assert_array_almost_equal(pca.singular_values_, [3.142, 2.718, 1.0], 14)
assert_array_almost_equal(ipca.singular_values_, [3.142, 2.718, 1.0], 14)
def test_compute_sample_weight_more_than_32():
# Non-regression smoke test for #12146
y = np.arange(50) # more than 32 distinct classes
indices = np.arange(50) # use subsampling
weight = compute_sample_weight("balanced", y, indices=indices)
assert_array_almost_equal(weight, np.ones(y.shape[0]))
def test_sag_multiclass_computed_correctly():
"""tests if the multiclass classifier is computed correctly"""
alpha = 0.1
n_samples = 20
tol = 0.00001
max_iter = 40
fit_intercept = True
X, y = make_blobs(n_samples=n_samples,
centers=3,
random_state=0,
cluster_std=0.1)
step_size = get_step_size(X, alpha, fit_intercept, classification=True)
classes = np.unique(y)
clf1 = LogisticRegression(
solver="sag",
C=1.0 / alpha / n_samples,
max_iter=max_iter,
tol=tol,
random_state=77,
fit_intercept=fit_intercept,
multi_class="ovr",
)
clf2 = clone(clf1)
clf1.fit(X, y)
clf2.fit(sp.csr_matrix(X), y)
coef1 = []
intercept1 = []
coef2 = []
intercept2 = []
for cl in classes:
y_encoded = np.ones(n_samples)
y_encoded[y != cl] = -1
spweights1, spintercept1 = sag_sparse(
X,
y_encoded,
step_size,
alpha,
dloss=log_dloss,
n_iter=max_iter,
fit_intercept=fit_intercept,
)
spweights2, spintercept2 = sag_sparse(
X,
y_encoded,
step_size,
alpha,
dloss=log_dloss,
n_iter=max_iter,
sparse=True,
fit_intercept=fit_intercept,
)
coef1.append(spweights1)
intercept1.append(spintercept1)
coef2.append(spweights2)
intercept2.append(spintercept2)
coef1 = np.vstack(coef1)
intercept1 = np.array(intercept1)
coef2 = np.vstack(coef2)
intercept2 = np.array(intercept2)
for i, cl in enumerate(classes):
assert_array_almost_equal(clf1.coef_[i].ravel(),
coef1[i].ravel(),
decimal=2)
assert_almost_equal(clf1.intercept_[i], intercept1[i], decimal=1)
assert_array_almost_equal(clf2.coef_[i].ravel(),
coef2[i].ravel(),
decimal=2)
assert_almost_equal(clf2.intercept_[i], intercept2[i], decimal=1)
def test_compute_sample_weight():
# Test (and demo) compute_sample_weight.
# Test with balanced classes
y = np.asarray([1, 1, 1, 2, 2, 2])
sample_weight = compute_sample_weight("balanced", y)
assert_array_almost_equal(sample_weight, [1.0, 1.0, 1.0, 1.0, 1.0, 1.0])
# Test with user-defined weights
sample_weight = compute_sample_weight({1: 2, 2: 1}, y)
assert_array_almost_equal(sample_weight, [2.0, 2.0, 2.0, 1.0, 1.0, 1.0])
# Test with column vector of balanced classes
y = np.asarray([[1], [1], [1], [2], [2], [2]])
sample_weight = compute_sample_weight("balanced", y)
assert_array_almost_equal(sample_weight, [1.0, 1.0, 1.0, 1.0, 1.0, 1.0])
# Test with unbalanced classes
y = np.asarray([1, 1, 1, 2, 2, 2, 3])
sample_weight = compute_sample_weight("balanced", y)
expected_balanced = np.array(
[0.7777, 0.7777, 0.7777, 0.7777, 0.7777, 0.7777, 2.3333]
)
assert_array_almost_equal(sample_weight, expected_balanced, decimal=4)
# Test with `None` weights
sample_weight = compute_sample_weight(None, y)
assert_array_almost_equal(sample_weight, [1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0])
# Test with multi-output of balanced classes
y = np.asarray([[1, 0], [1, 0], [1, 0], [2, 1], [2, 1], [2, 1]])
sample_weight = compute_sample_weight("balanced", y)
assert_array_almost_equal(sample_weight, [1.0, 1.0, 1.0, 1.0, 1.0, 1.0])
# Test with multi-output with user-defined weights
y = np.asarray([[1, 0], [1, 0], [1, 0], [2, 1], [2, 1], [2, 1]])
sample_weight = compute_sample_weight([{1: 2, 2: 1}, {0: 1, 1: 2}], y)
assert_array_almost_equal(sample_weight, [2.0, 2.0, 2.0, 2.0, 2.0, 2.0])
# Test with multi-output of unbalanced classes
y = np.asarray([[1, 0], [1, 0], [1, 0], [2, 1], [2, 1], [2, 1], [3, -1]])
sample_weight = compute_sample_weight("balanced", y)
assert_array_almost_equal(sample_weight, expected_balanced**2, decimal=3)
def test_compute_sample_weight_with_subsample():
# Test compute_sample_weight with subsamples specified.
# Test with balanced classes and all samples present
y = np.asarray([1, 1, 1, 2, 2, 2])
sample_weight = compute_sample_weight("balanced", y, indices=range(6))
assert_array_almost_equal(sample_weight, [1.0, 1.0, 1.0, 1.0, 1.0, 1.0])
# Test with column vector of balanced classes and all samples present
y = np.asarray([[1], [1], [1], [2], [2], [2]])
sample_weight = compute_sample_weight("balanced", y, indices=range(6))
assert_array_almost_equal(sample_weight, [1.0, 1.0, 1.0, 1.0, 1.0, 1.0])
# Test with a subsample
y = np.asarray([1, 1, 1, 2, 2, 2])
sample_weight = compute_sample_weight("balanced", y, indices=range(4))
assert_array_almost_equal(sample_weight, [2.0 / 3, 2.0 / 3, 2.0 / 3, 2.0, 2.0, 2.0])
# Test with a bootstrap subsample
y = np.asarray([1, 1, 1, 2, 2, 2])
sample_weight = compute_sample_weight("balanced", y, indices=[0, 1, 1, 2, 2, 3])
expected_balanced = np.asarray([0.6, 0.6, 0.6, 3.0, 3.0, 3.0])
assert_array_almost_equal(sample_weight, expected_balanced)
# Test with a bootstrap subsample for multi-output
y = np.asarray([[1, 0], [1, 0], [1, 0], [2, 1], [2, 1], [2, 1]])
sample_weight = compute_sample_weight("balanced", y, indices=[0, 1, 1, 2, 2, 3])
assert_array_almost_equal(sample_weight, expected_balanced**2)
# Test with a missing class
y = np.asarray([1, 1, 1, 2, 2, 2, 3])
sample_weight = compute_sample_weight("balanced", y, indices=range(6))
assert_array_almost_equal(sample_weight, [1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 0.0])
# Test with a missing class for multi-output
y = np.asarray([[1, 0], [1, 0], [1, 0], [2, 1], [2, 1], [2, 1], [2, 2]])
sample_weight = compute_sample_weight("balanced", y, indices=range(6))
assert_array_almost_equal(sample_weight, [1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 0.0])
def test_additive_chi2_sampler():
# test that AdditiveChi2Sampler approximates kernel on random data
# compute exact kernel
# abbreviations for easier formula
X_ = X[:, np.newaxis, :]
Y_ = Y[np.newaxis, :, :]
large_kernel = 2 * X_ * Y_ / (X_ + Y_)
# reduce to n_samples_x x n_samples_y by summing over features
kernel = large_kernel.sum(axis=2)
# approximate kernel mapping
transform = AdditiveChi2Sampler(sample_steps=3)
X_trans = transform.fit_transform(X)
Y_trans = transform.transform(Y)
kernel_approx = np.dot(X_trans, Y_trans.T)
assert_array_almost_equal(kernel, kernel_approx, 1)
X_sp_trans = transform.fit_transform(csr_matrix(X))
Y_sp_trans = transform.transform(csr_matrix(Y))
assert_array_equal(X_trans, X_sp_trans.A)
assert_array_equal(Y_trans, Y_sp_trans.A)
# test error is raised on negative input
Y_neg = Y.copy()
Y_neg[0, 0] = -1
msg = "Negative values in data passed to"
with pytest.raises(ValueError, match=msg):
transform.transform(Y_neg)
# test error on invalid sample_steps
transform = AdditiveChi2Sampler(sample_steps=4)
msg = re.escape(
"If sample_steps is not in [1, 2, 3], you need to provide sample_interval"
)
with pytest.raises(ValueError, match=msg):
transform.fit(X)
# test that the sample interval is set correctly
sample_steps_available = [1, 2, 3]
for sample_steps in sample_steps_available:
# test that the sample_interval is initialized correctly
transform = AdditiveChi2Sampler(sample_steps=sample_steps)
assert transform.sample_interval is None
# test that the sample_interval is changed in the fit method
transform.fit(X)
assert transform.sample_interval_ is not None
# test that the sample_interval is set correctly
sample_interval = 0.3
transform = AdditiveChi2Sampler(sample_steps=4, sample_interval=sample_interval)
assert transform.sample_interval == sample_interval
transform.fit(X)
assert transform.sample_interval_ == sample_interval
def test_sparse_classification():
# Check classification for various parameter settings on sparse input.
class CustomSVC(SVC):
"""SVC variant that records the nature of the training set"""
def fit(self, X, y):
super().fit(X, y)
self.data_type_ = type(X)
return self
rng = check_random_state(0)
X_train, X_test, y_train, y_test = train_test_split(iris.data,
iris.target,
random_state=rng)
parameter_sets = [
{
"max_samples": 0.5,
"max_features": 2,
"bootstrap": True,
"bootstrap_features": True
},
{
"max_samples": 1.0,
"max_features": 4,
"bootstrap": True,
"bootstrap_features": True
},
{
"max_features": 2,
"bootstrap": False,
"bootstrap_features": True
},
{
"max_samples": 0.5,
"bootstrap": True,
"bootstrap_features": False
},
]
for sparse_format in [csc_matrix, csr_matrix]:
X_train_sparse = sparse_format(X_train)
X_test_sparse = sparse_format(X_test)
for params in parameter_sets:
for f in [
'predict', 'predict_proba', 'predict_log_proba',
'decision_function'
]:
# Trained on sparse format
sparse_classifier = BaggingClassifier(
base_estimator=CustomSVC(decision_function_shape='ovr'),
random_state=1,
**params).fit(X_train_sparse, y_train)
sparse_results = getattr(sparse_classifier, f)(X_test_sparse)
# Trained on dense format
dense_classifier = BaggingClassifier(
base_estimator=CustomSVC(decision_function_shape='ovr'),
random_state=1,
**params).fit(X_train, y_train)
dense_results = getattr(dense_classifier, f)(X_test)
assert_array_almost_equal(sparse_results, dense_results)
sparse_type = type(X_train_sparse)
types = [i.data_type_ for i in sparse_classifier.estimators_]
assert all([t == sparse_type for t in types])
def test_nmf_multiplicative_update_sparse():
# Compare sparse and dense input in multiplicative update NMF
# Also test continuity of the results with respect to beta_loss parameter
n_samples = 20
n_features = 10
n_components = 5
alpha = 0.1
l1_ratio = 0.5
n_iter = 20
# initialization
rng = np.random.mtrand.RandomState(1337)
X = rng.randn(n_samples, n_features)
X = np.abs(X)
X_csr = sp.csr_matrix(X)
W0, H0 = nmf._initialize_nmf(X,
n_components,
init='random',
random_state=42)
for beta_loss in (-1.2, 0, 0.2, 1., 2., 2.5):
# Reference with dense array X
W, H = W0.copy(), H0.copy()
W1, H1, _ = non_negative_factorization(X,
W,
H,
n_components,
init='custom',
update_H=True,
solver='mu',
beta_loss=beta_loss,
max_iter=n_iter,
alpha=alpha,
l1_ratio=l1_ratio,
regularization='both',
random_state=42)
# Compare with sparse X
W, H = W0.copy(), H0.copy()
W2, H2, _ = non_negative_factorization(X_csr,
W,
H,
n_components,
init='custom',
update_H=True,
solver='mu',
beta_loss=beta_loss,
max_iter=n_iter,
alpha=alpha,
l1_ratio=l1_ratio,
regularization='both',
random_state=42)
assert_array_almost_equal(W1, W2, decimal=7)
assert_array_almost_equal(H1, H2, decimal=7)
# Compare with almost same beta_loss, since some values have a specific
# behavior, but the results should be continuous w.r.t beta_loss
beta_loss -= 1.e-5
W, H = W0.copy(), H0.copy()
W3, H3, _ = non_negative_factorization(X_csr,
W,
H,
n_components,
init='custom',
update_H=True,
solver='mu',
beta_loss=beta_loss,
max_iter=n_iter,
alpha=alpha,
l1_ratio=l1_ratio,
regularization='both',
random_state=42)
assert_array_almost_equal(W1, W3, decimal=4)
assert_array_almost_equal(H1, H3, decimal=4)
def test_incremental_variance_numerical_stability():
# Test Youngs and Cramer incremental variance formulas.
def np_var(A):
return A.var(axis=0)
# Naive one pass variance computation - not numerically stable
# https://en.wikipedia.org/wiki/Algorithms_for_calculating_variance
def one_pass_var(X):
n = X.shape[0]
exp_x2 = (X ** 2).sum(axis=0) / n
expx_2 = (X.sum(axis=0) / n) ** 2
return exp_x2 - expx_2
# Two-pass algorithm, stable.
# We use it as a benchmark. It is not an online algorithm
# https://en.wikipedia.org/wiki/Algorithms_for_calculating_variance#Two-pass_algorithm
def two_pass_var(X):
mean = X.mean(axis=0)
Y = X.copy()
return np.mean((Y - mean)**2, axis=0)
# Naive online implementation
# https://en.wikipedia.org/wiki/Algorithms_for_calculating_variance#Online_algorithm
# This works only for chunks for size 1
def naive_mean_variance_update(x, last_mean, last_variance,
last_sample_count):
updated_sample_count = (last_sample_count + 1)
samples_ratio = last_sample_count / float(updated_sample_count)
updated_mean = x / updated_sample_count + last_mean * samples_ratio
updated_variance = last_variance * samples_ratio + \
(x - last_mean) * (x - updated_mean) / updated_sample_count
return updated_mean, updated_variance, updated_sample_count
# We want to show a case when one_pass_var has error > 1e-3 while
# _batch_mean_variance_update has less.
tol = 200
n_features = 2
n_samples = 10000
x1 = np.array(1e8, dtype=np.float64)
x2 = np.log(1e-5, dtype=np.float64)
A0 = np.full((n_samples // 2, n_features), x1, dtype=np.float64)
A1 = np.full((n_samples // 2, n_features), x2, dtype=np.float64)
A = np.vstack((A0, A1))
# Naive one pass var: >tol (=1063)
assert np.abs(np_var(A) - one_pass_var(A)).max() > tol
# Starting point for online algorithms: after A0
# Naive implementation: >tol (436)
mean, var, n = A0[0, :], np.zeros(n_features), n_samples // 2
for i in range(A1.shape[0]):
mean, var, n = \
naive_mean_variance_update(A1[i, :], mean, var, n)
assert n == A.shape[0]
# the mean is also slightly unstable
assert np.abs(A.mean(axis=0) - mean).max() > 1e-6
assert np.abs(np_var(A) - var).max() > tol
# Robust implementation: <tol (177)
mean, var = A0[0, :], np.zeros(n_features)
n = np.full(n_features, n_samples // 2, dtype=np.int32)
for i in range(A1.shape[0]):
mean, var, n = \
_incremental_mean_and_var(A1[i, :].reshape((1, A1.shape[1])),
mean, var, n)
assert_array_equal(n, A.shape[0])
assert_array_almost_equal(A.mean(axis=0), mean)
assert tol > np.abs(np_var(A) - var).max()
def test_binary_classifier_class_weight():
"""tests binary classifier with classweights for each class"""
alpha = 0.1
n_samples = 50
n_iter = 20
tol = 0.00001
fit_intercept = True
X, y = make_blobs(n_samples=n_samples,
centers=2,
random_state=10,
cluster_std=0.1)
step_size = get_step_size(X, alpha, fit_intercept, classification=True)
classes = np.unique(y)
y_tmp = np.ones(n_samples)
y_tmp[y != classes[1]] = -1
y = y_tmp
class_weight = {1: 0.45, -1: 0.55}
clf1 = LogisticRegression(
solver="sag",
C=1.0 / alpha / n_samples,
max_iter=n_iter,
tol=tol,
random_state=77,
fit_intercept=fit_intercept,
multi_class="ovr",
class_weight=class_weight,
)
clf2 = clone(clf1)
clf1.fit(X, y)
clf2.fit(sp.csr_matrix(X), y)
le = LabelEncoder()
class_weight_ = compute_class_weight(class_weight,
classes=np.unique(y),
y=y)
sample_weight = class_weight_[le.fit_transform(y)]
spweights, spintercept = sag_sparse(
X,
y,
step_size,
alpha,
n_iter=n_iter,
dloss=log_dloss,
sample_weight=sample_weight,
fit_intercept=fit_intercept,
)
spweights2, spintercept2 = sag_sparse(
X,
y,
step_size,
alpha,
n_iter=n_iter,
dloss=log_dloss,
sparse=True,
sample_weight=sample_weight,
fit_intercept=fit_intercept,
)
assert_array_almost_equal(clf1.coef_.ravel(), spweights.ravel(), decimal=2)
assert_almost_equal(clf1.intercept_, spintercept, decimal=1)
assert_array_almost_equal(clf2.coef_.ravel(),
spweights2.ravel(),
decimal=2)
assert_almost_equal(clf2.intercept_, spintercept2, decimal=1)
def test_softmax():
rng = np.random.RandomState(0)
X = rng.randn(3, 5)
exp_X = np.exp(X)
sum_exp_X = np.sum(exp_X, axis=1).reshape((-1, 1))
assert_array_almost_equal(softmax(X), exp_X / sum_exp_X)
def test_multiclass_classifier_class_weight():
"""tests multiclass with classweights for each class"""
alpha = 0.1
n_samples = 20
tol = 0.00001
max_iter = 50
class_weight = {0: 0.45, 1: 0.55, 2: 0.75}
fit_intercept = True
X, y = make_blobs(n_samples=n_samples,
centers=3,
random_state=0,
cluster_std=0.1)
step_size = get_step_size(X, alpha, fit_intercept, classification=True)
classes = np.unique(y)
clf1 = LogisticRegression(
solver="sag",
C=1.0 / alpha / n_samples,
max_iter=max_iter,
tol=tol,
random_state=77,
fit_intercept=fit_intercept,
multi_class="ovr",
class_weight=class_weight,
)
clf2 = clone(clf1)
clf1.fit(X, y)
clf2.fit(sp.csr_matrix(X), y)
le = LabelEncoder()
class_weight_ = compute_class_weight(class_weight,
classes=np.unique(y),
y=y)
sample_weight = class_weight_[le.fit_transform(y)]
coef1 = []
intercept1 = []
coef2 = []
intercept2 = []
for cl in classes:
y_encoded = np.ones(n_samples)
y_encoded[y != cl] = -1
spweights1, spintercept1 = sag_sparse(
X,
y_encoded,
step_size,
alpha,
n_iter=max_iter,
dloss=log_dloss,
sample_weight=sample_weight,
)
spweights2, spintercept2 = sag_sparse(
X,
y_encoded,
step_size,
alpha,
n_iter=max_iter,
dloss=log_dloss,
sample_weight=sample_weight,
sparse=True,
)
coef1.append(spweights1)
intercept1.append(spintercept1)
coef2.append(spweights2)
intercept2.append(spintercept2)
coef1 = np.vstack(coef1)
intercept1 = np.array(intercept1)
coef2 = np.vstack(coef2)
intercept2 = np.array(intercept2)
for i, cl in enumerate(classes):
assert_array_almost_equal(clf1.coef_[i].ravel(),
coef1[i].ravel(),
decimal=2)
assert_almost_equal(clf1.intercept_[i], intercept1[i], decimal=1)
assert_array_almost_equal(clf2.coef_[i].ravel(),
coef2[i].ravel(),
decimal=2)
assert_almost_equal(clf2.intercept_[i], intercept2[i], decimal=1)
def test_enet_toy():
# Test ElasticNet for various parameters of alpha and l1_ratio.
# Actually, the parameters alpha = 0 should not be allowed. However,
# we test it as a border case.
# ElasticNet is tested with and without precomputed Gram matrix
X = np.array([[-1.], [0.], [1.]])
Y = [-1, 0, 1] # just a straight line
T = [[2.], [3.], [4.]] # test sample
# this should be the same as lasso
clf = ElasticNet(alpha=1e-8, l1_ratio=1.0)
clf.fit(X, Y)
pred = clf.predict(T)
assert_array_almost_equal(clf.coef_, [1])
assert_array_almost_equal(pred, [2, 3, 4])
assert_almost_equal(clf.dual_gap_, 0)
clf = ElasticNet(alpha=0.5, l1_ratio=0.3, max_iter=100, precompute=False)
clf.fit(X, Y)
pred = clf.predict(T)
assert_array_almost_equal(clf.coef_, [0.50819], decimal=3)
assert_array_almost_equal(pred, [1.0163, 1.5245, 2.0327], decimal=3)
assert_almost_equal(clf.dual_gap_, 0)
clf.set_params(max_iter=100, precompute=True)
clf.fit(X, Y) # with Gram
pred = clf.predict(T)
assert_array_almost_equal(clf.coef_, [0.50819], decimal=3)
assert_array_almost_equal(pred, [1.0163, 1.5245, 2.0327], decimal=3)
assert_almost_equal(clf.dual_gap_, 0)
clf.set_params(max_iter=100, precompute=np.dot(X.T, X))
clf.fit(X, Y) # with Gram
pred = clf.predict(T)
assert_array_almost_equal(clf.coef_, [0.50819], decimal=3)
assert_array_almost_equal(pred, [1.0163, 1.5245, 2.0327], decimal=3)
assert_almost_equal(clf.dual_gap_, 0)
clf = ElasticNet(alpha=0.5, l1_ratio=0.5)
clf.fit(X, Y)
pred = clf.predict(T)
assert_array_almost_equal(clf.coef_, [0.45454], 3)
assert_array_almost_equal(pred, [0.9090, 1.3636, 1.8181], 3)
assert_almost_equal(clf.dual_gap_, 0)
def test_explained_variance(setup):
# Check that PCA output has unit-variance
rng = np.random.RandomState(0)
n_samples = 100
n_features = 80
X = mt.tensor(rng.randn(n_samples, n_features))
pca = PCA(n_components=2, svd_solver='full').fit(X)
rpca = PCA(n_components=2, svd_solver='randomized', random_state=42).fit(X)
assert_array_almost_equal(pca.explained_variance_.to_numpy(),
rpca.explained_variance_.to_numpy(), 1)
assert_array_almost_equal(pca.explained_variance_ratio_.to_numpy(),
rpca.explained_variance_ratio_.to_numpy(), 1)
# compare to empirical variances
expected_result = np.linalg.eig(np.cov(X.to_numpy(), rowvar=False))[0]
expected_result = sorted(expected_result, reverse=True)[:2]
X_pca = pca.transform(X)
assert_array_almost_equal(pca.explained_variance_.to_numpy(),
mt.var(X_pca, ddof=1, axis=0).to_numpy())
assert_array_almost_equal(pca.explained_variance_.to_numpy(), expected_result)
X_rpca = rpca.transform(X)
assert_array_almost_equal(rpca.explained_variance_.to_numpy(),
mt.var(X_rpca, ddof=1, axis=0).to_numpy(),
decimal=1)
assert_array_almost_equal(rpca.explained_variance_.to_numpy(),
expected_result, decimal=1)
# Same with correlated data
X = datasets.make_classification(n_samples, n_features,
n_informative=n_features-2,
random_state=rng)[0]
X = mt.tensor(X)
pca = PCA(n_components=2).fit(X)
rpca = PCA(n_components=2, svd_solver='randomized',
random_state=rng).fit(X)
assert_array_almost_equal(pca.explained_variance_ratio_.to_numpy(),
rpca.explained_variance_ratio_.to_numpy(), 5)
def test_sparse_regression():
# Check regression for various parameter settings on sparse input.
rng = check_random_state(0)
X_train, X_test, y_train, y_test = train_test_split(diabetes.data[:50],
diabetes.target[:50],
random_state=rng)
class CustomSVR(SVR):
"""SVC variant that records the nature of the training set"""
def fit(self, X, y):
super().fit(X, y)
self.data_type_ = type(X)
return self
parameter_sets = [
{
"max_samples": 0.5,
"max_features": 2,
"bootstrap": True,
"bootstrap_features": True
},
{
"max_samples": 1.0,
"max_features": 4,
"bootstrap": True,
"bootstrap_features": True
},
{
"max_features": 2,
"bootstrap": False,
"bootstrap_features": True
},
{
"max_samples": 0.5,
"bootstrap": True,
"bootstrap_features": False
},
]
for sparse_format in [csc_matrix, csr_matrix]:
X_train_sparse = sparse_format(X_train)
X_test_sparse = sparse_format(X_test)
for params in parameter_sets:
# Trained on sparse format
sparse_classifier = BaggingRegressor(base_estimator=CustomSVR(),
random_state=1,
**params).fit(
X_train_sparse, y_train)
sparse_results = sparse_classifier.predict(X_test_sparse)
# Trained on dense format
dense_results = BaggingRegressor(base_estimator=CustomSVR(),
random_state=1,
**params).fit(
X_train,
y_train).predict(X_test)
sparse_type = type(X_train_sparse)
types = [i.data_type_ for i in sparse_classifier.estimators_]
assert_array_almost_equal(sparse_results, dense_results)
assert all([t == sparse_type for t in types])
assert_array_almost_equal(sparse_results, dense_results)
def test_singular_values(setup):
# Check that the PCA output has the correct singular values
rng = np.random.RandomState(0)
n_samples = 100
n_features = 80
X = mt.tensor(rng.randn(n_samples, n_features))
pca = PCA(n_components=2, svd_solver='full',
random_state=rng).fit(X)
rpca = PCA(n_components=2, svd_solver='randomized',
random_state=rng).fit(X)
assert_array_almost_equal(pca.singular_values_.fetch(),
rpca.singular_values_.fetch(), 1)
# Compare to the Frobenius norm
X_pca = pca.transform(X)
X_rpca = rpca.transform(X)
assert_array_almost_equal(mt.sum(pca.singular_values_**2.0).to_numpy(),
(mt.linalg.norm(X_pca, "fro")**2.0).to_numpy(), 12)
assert_array_almost_equal(mt.sum(rpca.singular_values_**2.0).to_numpy(),
(mt.linalg.norm(X_rpca, "fro")**2.0).to_numpy(), 0)
# Compare to the 2-norms of the score vectors
assert_array_almost_equal(pca.singular_values_.fetch(),
mt.sqrt(mt.sum(X_pca**2.0, axis=0)).to_numpy(), 12)
assert_array_almost_equal(rpca.singular_values_.fetch(),
mt.sqrt(mt.sum(X_rpca**2.0, axis=0)).to_numpy(), 2)
# Set the singular values and see what we get back
rng = np.random.RandomState(0)
n_samples = 100
n_features = 110
X = mt.tensor(rng.randn(n_samples, n_features))
pca = PCA(n_components=3, svd_solver='full', random_state=rng)
rpca = PCA(n_components=3, svd_solver='randomized', random_state=rng)
X_pca = pca.fit_transform(X)
X_pca /= mt.sqrt(mt.sum(X_pca**2.0, axis=0))
X_pca[:, 0] *= 3.142
X_pca[:, 1] *= 2.718
X_hat = mt.dot(X_pca, pca.components_)
pca.fit(X_hat)
rpca.fit(X_hat)
assert_array_almost_equal(pca.singular_values_.fetch(), [3.142, 2.718, 1.0], 14)
assert_array_almost_equal(rpca.singular_values_.fetch(), [3.142, 2.718, 1.0], 14)
def test_ward_linkage_tree_return_distance():
# Test return_distance option on linkage and ward trees
# test that return_distance when set true, gives same
# output on both structured and unstructured clustering.
n, p = 10, 5
rng = np.random.RandomState(0)
connectivity = np.ones((n, n))
for i in range(5):
X = .1 * rng.normal(size=(n, p))
X -= 4. * np.arange(n)[:, np.newaxis]
X -= X.mean(axis=1)[:, np.newaxis]
out_unstructured = ward_tree(X, return_distance=True)
out_structured = ward_tree(X,
connectivity=connectivity,
return_distance=True)
# get children
children_unstructured = out_unstructured[0]
children_structured = out_structured[0]
# check if we got the same clusters
assert_array_equal(children_unstructured, children_structured)
# check if the distances are the same
dist_unstructured = out_unstructured[-1]
dist_structured = out_structured[-1]
assert_array_almost_equal(dist_unstructured, dist_structured)
for linkage in ['average', 'complete', 'single']:
structured_items = linkage_tree(X,
connectivity=connectivity,
linkage=linkage,
return_distance=True)[-1]
unstructured_items = linkage_tree(X,
linkage=linkage,
return_distance=True)[-1]
structured_dist = structured_items[-1]
unstructured_dist = unstructured_items[-1]
structured_children = structured_items[0]
unstructured_children = unstructured_items[0]
assert_array_almost_equal(structured_dist, unstructured_dist)
assert_array_almost_equal(structured_children,
unstructured_children)
# test on the following dataset where we know the truth
# taken from scipy/cluster/tests/hierarchy_test_data.py
X = np.array([[1.43054825, -7.5693489], [6.95887839, 6.82293382],
[2.87137846, -9.68248579], [7.87974764, -6.05485803],
[8.24018364, -6.09495602], [7.39020262, 8.54004355]])
# truth
linkage_X_ward = np.array([[3., 4., 0.36265956, 2.],
[1., 5., 1.77045373, 2.],
[0., 2., 2.55760419, 2.],
[6., 8., 9.10208346, 4.],
[7., 9., 24.7784379, 6.]])
linkage_X_complete = np.array([[3., 4., 0.36265956, 2.],
[1., 5., 1.77045373, 2.],
[0., 2., 2.55760419, 2.],
[6., 8., 6.96742194, 4.],
[7., 9., 18.77445997, 6.]])
linkage_X_average = np.array([[3., 4., 0.36265956, 2.],
[1., 5., 1.77045373, 2.],
[0., 2., 2.55760419, 2.],
[6., 8., 6.55832839, 4.],
[7., 9., 15.44089605, 6.]])
n_samples, n_features = np.shape(X)
connectivity_X = np.ones((n_samples, n_samples))
out_X_unstructured = ward_tree(X, return_distance=True)
out_X_structured = ward_tree(X,
connectivity=connectivity_X,
return_distance=True)
# check that the labels are the same
assert_array_equal(linkage_X_ward[:, :2], out_X_unstructured[0])
assert_array_equal(linkage_X_ward[:, :2], out_X_structured[0])
# check that the distances are correct
assert_array_almost_equal(linkage_X_ward[:, 2], out_X_unstructured[4])
assert_array_almost_equal(linkage_X_ward[:, 2], out_X_structured[4])
linkage_options = ['complete', 'average', 'single']
X_linkage_truth = [linkage_X_complete, linkage_X_average]
for (linkage, X_truth) in zip(linkage_options, X_linkage_truth):
out_X_unstructured = linkage_tree(X,
return_distance=True,
linkage=linkage)
out_X_structured = linkage_tree(X,
connectivity=connectivity_X,
linkage=linkage,
return_distance=True)
# check that the labels are the same
assert_array_equal(X_truth[:, :2], out_X_unstructured[0])
assert_array_equal(X_truth[:, :2], out_X_structured[0])
# check that the distances are correct
assert_array_almost_equal(X_truth[:, 2], out_X_unstructured[4])
assert_array_almost_equal(X_truth[:, 2], out_X_structured[4])
def test_dump():
X_sparse, y_dense = load_svmlight_file(datafile)
X_dense = X_sparse.toarray()
y_sparse = sp.csr_matrix(y_dense)
# slicing a csr_matrix can unsort its .indices, so test that we sort
# those correctly
X_sliced = X_sparse[np.arange(X_sparse.shape[0])]
y_sliced = y_sparse[np.arange(y_sparse.shape[0])]
for X in (X_sparse, X_dense, X_sliced):
for y in (y_sparse, y_dense, y_sliced):
for zero_based in (True, False):
for dtype in [np.float32, np.float64, np.int32, np.int64]:
f = BytesIO()
# we need to pass a comment to get the version info in;
# LibSVM doesn't grok comments so they're not put in by
# default anymore.
if (sp.issparse(y) and y.shape[0] == 1):
# make sure y's shape is: (n_samples, n_labels)
# when it is sparse
y = y.T
# Note: with dtype=np.int32 we are performing unsafe casts,
# where X.astype(dtype) overflows. The result is
# then platform dependent and X_dense.astype(dtype) may be
# different from X_sparse.astype(dtype).asarray().
X_input = X.astype(dtype)
dump_svmlight_file(X_input, y, f, comment="test",
zero_based=zero_based)
f.seek(0)
comment = f.readline()
comment = str(comment, "utf-8")
assert "scikit-learn %s" % sklearn.__version__ in comment
comment = f.readline()
comment = str(comment, "utf-8")
assert ["one", "zero"][zero_based] + "-based" in comment
X2, y2 = load_svmlight_file(f, dtype=dtype,
zero_based=zero_based)
assert X2.dtype == dtype
assert_array_equal(X2.sorted_indices().indices, X2.indices)
X2_dense = X2.toarray()
if sp.issparse(X_input):
X_input_dense = X_input.toarray()
else:
X_input_dense = X_input
if dtype == np.float32:
# allow a rounding error at the last decimal place
assert_array_almost_equal(
X_input_dense, X2_dense, 4)
assert_array_almost_equal(
y_dense.astype(dtype, copy=False), y2, 4)
else:
# allow a rounding error at the last decimal place
assert_array_almost_equal(
X_input_dense, X2_dense, 15)
assert_array_almost_equal(
y_dense.astype(dtype, copy=False), y2, 15)
def test_set_estimator_none(drop):
"""VotingClassifier set_params should be able to set estimators as None or
drop"""
# Test predict
clf1 = LogisticRegression(random_state=123)
clf2 = RandomForestClassifier(n_estimators=10, random_state=123)
clf3 = GaussianNB()
eclf1 = VotingClassifier(estimators=[('lr', clf1), ('rf', clf2),
('nb', clf3)],
voting='hard',
weights=[1, 0, 0.5]).fit(X, y)
eclf2 = VotingClassifier(estimators=[('lr', clf1), ('rf', clf2),
('nb', clf3)],
voting='hard',
weights=[1, 1, 0.5])
with pytest.warns(None) as record:
eclf2.set_params(rf=drop).fit(X, y)
assert record if drop is None else not record
assert_array_equal(eclf1.predict(X), eclf2.predict(X))
assert dict(eclf2.estimators)["rf"] is drop
assert len(eclf2.estimators_) == 2
assert all(
isinstance(est, (LogisticRegression, GaussianNB))
for est in eclf2.estimators_)
assert eclf2.get_params()["rf"] is drop
eclf1.set_params(voting='soft').fit(X, y)
with pytest.warns(None) as record:
eclf2.set_params(voting='soft').fit(X, y)
assert record if drop is None else not record
assert_array_equal(eclf1.predict(X), eclf2.predict(X))
assert_array_almost_equal(eclf1.predict_proba(X), eclf2.predict_proba(X))
msg = 'All estimators are dropped. At least one is required'
with pytest.warns(None) as record:
with pytest.raises(ValueError, match=msg):
eclf2.set_params(lr=drop, rf=drop, nb=drop).fit(X, y)
assert record if drop is None else not record
# Test soft voting transform
X1 = np.array([[1], [2]])
y1 = np.array([1, 2])
eclf1 = VotingClassifier(estimators=[('rf', clf2), ('nb', clf3)],
voting='soft',
weights=[0, 0.5],
flatten_transform=False).fit(X1, y1)
eclf2 = VotingClassifier(estimators=[('rf', clf2), ('nb', clf3)],
voting='soft',
weights=[1, 0.5],
flatten_transform=False)
with pytest.warns(None) as record:
eclf2.set_params(rf=drop).fit(X1, y1)
assert record if drop is None else not record
assert_array_almost_equal(
eclf1.transform(X1),
np.array([[[0.7, 0.3], [0.3, 0.7]], [[1., 0.], [0., 1.]]]))
assert_array_almost_equal(eclf2.transform(X1),
np.array([[[1., 0.], [0., 1.]]]))
eclf1.set_params(voting='hard')
eclf2.set_params(voting='hard')
assert_array_equal(eclf1.transform(X1), np.array([[0, 0], [1, 1]]))
assert_array_equal(eclf2.transform(X1), np.array([[0], [1]]))
def test_minibatch_update_consistency():
# Check that dense and sparse minibatch update give the same results
rng = np.random.RandomState(42)
old_centers = centers + rng.normal(size=centers.shape)
new_centers = old_centers.copy()
new_centers_csr = old_centers.copy()
weight_sums = np.zeros(new_centers.shape[0], dtype=np.double)
weight_sums_csr = np.zeros(new_centers.shape[0], dtype=np.double)
x_squared_norms = (X ** 2).sum(axis=1)
x_squared_norms_csr = row_norms(X_csr, squared=True)
buffer = np.zeros(centers.shape[1], dtype=np.double)
buffer_csr = np.zeros(centers.shape[1], dtype=np.double)
# extract a small minibatch
X_mb = X[:10]
X_mb_csr = X_csr[:10]
x_mb_squared_norms = x_squared_norms[:10]
x_mb_squared_norms_csr = x_squared_norms_csr[:10]
sample_weight_mb = np.ones(X_mb.shape[0], dtype=np.double)
# step 1: compute the dense minibatch update
old_inertia, incremental_diff = _mini_batch_step(
X_mb, sample_weight_mb, x_mb_squared_norms, new_centers, weight_sums,
buffer, 1, None, random_reassign=False)
assert old_inertia > 0.0
# compute the new inertia on the same batch to check that it decreased
labels, new_inertia = _labels_inertia(
X_mb, sample_weight_mb, x_mb_squared_norms, new_centers)
assert new_inertia > 0.0
assert new_inertia < old_inertia
# check that the incremental difference computation is matching the
# final observed value
effective_diff = np.sum((new_centers - old_centers) ** 2)
assert_almost_equal(incremental_diff, effective_diff)
# step 2: compute the sparse minibatch update
old_inertia_csr, incremental_diff_csr = _mini_batch_step(
X_mb_csr, sample_weight_mb, x_mb_squared_norms_csr, new_centers_csr,
weight_sums_csr, buffer_csr, 1, None, random_reassign=False)
assert old_inertia_csr > 0.0
# compute the new inertia on the same batch to check that it decreased
labels_csr, new_inertia_csr = _labels_inertia(
X_mb_csr, sample_weight_mb, x_mb_squared_norms_csr, new_centers_csr)
assert new_inertia_csr > 0.0
assert new_inertia_csr < old_inertia_csr
# check that the incremental difference computation is matching the
# final observed value
effective_diff = np.sum((new_centers_csr - old_centers) ** 2)
assert_almost_equal(incremental_diff_csr, effective_diff)
# step 3: check that sparse and dense updates lead to the same results
assert_array_equal(labels, labels_csr)
assert_array_almost_equal(new_centers, new_centers_csr)
assert_almost_equal(incremental_diff, incremental_diff_csr)
assert_almost_equal(old_inertia, old_inertia_csr)
assert_almost_equal(new_inertia, new_inertia_csr)
def test_singular_matrix():
# Test when input is a singular matrix
X1 = np.array([[1, 1.], [1., 1.]])
y1 = np.array([1, 1])
_, _, coef_path = linear_model.lars_path(X1, y1)
assert_array_almost_equal(coef_path.T, [[0, 0], [1, 0]])
def test_fit_transform():
X1 = KMeans(n_clusters=3, random_state=51).fit(X).transform(X)
X2 = KMeans(n_clusters=3, random_state=51).fit_transform(X)
assert_array_almost_equal(X1, X2)
def test_regression_multioutput_array():
y_true = [[1, 2], [2.5, -1], [4.5, 3], [5, 7]]
y_pred = [[1, 1], [2, -1], [5, 4], [5, 6.5]]
mse = mean_squared_error(y_true, y_pred, multioutput='raw_values')
mae = mean_absolute_error(y_true, y_pred, multioutput='raw_values')
mape = mean_absolute_percentage_error(y_true,
y_pred,
multioutput='raw_values')
r = r2_score(y_true, y_pred, multioutput='raw_values')
evs = explained_variance_score(y_true, y_pred, multioutput='raw_values')
assert_array_almost_equal(mse, [0.125, 0.5625], decimal=2)
assert_array_almost_equal(mae, [0.25, 0.625], decimal=2)
assert_array_almost_equal(mape, [0.0778, 0.2262], decimal=2)
assert_array_almost_equal(r, [0.95, 0.93], decimal=2)
assert_array_almost_equal(evs, [0.95, 0.93], decimal=2)
# mean_absolute_error and mean_squared_error are equal because
# it is a binary problem.
y_true = [[0, 0]] * 4
y_pred = [[1, 1]] * 4
mse = mean_squared_error(y_true, y_pred, multioutput='raw_values')
mae = mean_absolute_error(y_true, y_pred, multioutput='raw_values')
r = r2_score(y_true, y_pred, multioutput='raw_values')
assert_array_almost_equal(mse, [1., 1.], decimal=2)
assert_array_almost_equal(mae, [1., 1.], decimal=2)
assert_array_almost_equal(r, [0., 0.], decimal=2)
r = r2_score([[0, -1], [0, 1]], [[2, 2], [1, 1]], multioutput='raw_values')
assert_array_almost_equal(r, [0, -3.5], decimal=2)
assert np.mean(r) == r2_score([[0, -1], [0, 1]], [[2, 2], [1, 1]],
multioutput='uniform_average')
evs = explained_variance_score([[0, -1], [0, 1]], [[2, 2], [1, 1]],
multioutput='raw_values')
assert_array_almost_equal(evs, [0, -1.25], decimal=2)
# Checking for the condition in which both numerator and denominator is
# zero.
y_true = [[1, 3], [-1, 2]]
y_pred = [[1, 4], [-1, 1]]
r2 = r2_score(y_true, y_pred, multioutput='raw_values')
assert_array_almost_equal(r2, [1., -3.], decimal=2)
assert np.mean(r2) == r2_score(y_true,
y_pred,
multioutput='uniform_average')
evs = explained_variance_score(y_true, y_pred, multioutput='raw_values')
assert_array_almost_equal(evs, [1., -3.], decimal=2)
assert np.mean(evs) == explained_variance_score(y_true, y_pred)
# Handling msle separately as it does not accept negative inputs.
y_true = np.array([[0.5, 1], [1, 2], [7, 6]])
y_pred = np.array([[0.5, 2], [1, 2.5], [8, 8]])
msle = mean_squared_log_error(y_true, y_pred, multioutput='raw_values')
msle2 = mean_squared_error(np.log(1 + y_true),
np.log(1 + y_pred),
multioutput='raw_values')
assert_array_almost_equal(msle, msle2, decimal=2)
def test_auto_vs_cross(kernel):
# Auto-correlation and cross-correlation should be consistent.
K_auto = kernel(X)
K_cross = kernel(X, X)
print(K_auto, K_cross)
assert_array_almost_equal(K_auto, K_cross, 5)