def test_inner_product(self, n, power_affine, tangent_vec_a, tangent_vec_b,
base_point, expected):
metric = SPDMetricAffine(n, power_affine)
result = metric.inner_product(gs.array(tangent_vec_a),
gs.array(tangent_vec_b),
gs.array(base_point))
self.assertAllClose(result, expected)
def test_power_affine_inner_product(self):
base_point = gs.array([[1., 0., 0.], [0., 2.5, 1.5], [0., 1.5, 2.5]])
tangent_vec = gs.array([[2., 1., 1.], [1., .5, .5], [1., .5, .5]])
metric = SPDMetricAffine(3, power_affine=.5)
result = metric.inner_product(tangent_vec, tangent_vec, base_point)
expected = [[713 / 144]]
self.assertAllClose(result, expected)
def test_log_and_exp_power_affine(self):
base_point = gs.array([[5., 0., 0.], [0., 7., 2.], [0., 2., 8.]])
point = gs.array([[9., 0., 0.], [0., 5., 0.], [0., 0., 1.]])
metric = SPDMetricAffine(3, power_affine=.5)
log = metric.log(point, base_point)
result = metric.exp(log, base_point)
expected = point
self.assertAllClose(result, expected)
def test_log_and_exp_power_affine(self):
"""Test of SPDMetricAffine.log and exp methods with power!=1."""
base_point = gs.array([[5.0, 0.0, 0.0], [0.0, 7.0, 2.0], [0.0, 2.0, 8.0]])
point = gs.array([[9.0, 0.0, 0.0], [0.0, 5.0, 0.0], [0.0, 0.0, 1.0]])
metric = SPDMetricAffine(3, power_affine=0.5)
log = metric.log(point, base_point)
result = metric.exp(log, base_point)
expected = point
self.assertAllClose(result, expected)
def test_power_affine_inner_product(self):
"""Test of SPDMetricAffine.inner_product method."""
base_point = gs.array([[1.0, 0.0, 0.0], [0.0, 2.5, 1.5], [0.0, 1.5, 2.5]])
tangent_vec = gs.array([[2.0, 1.0, 1.0], [1.0, 0.5, 0.5], [1.0, 0.5, 0.5]])
metric = SPDMetricAffine(3, power_affine=0.5)
result = metric.inner_product(tangent_vec, tangent_vec, base_point)
expected = 713 / 144
self.assertAllClose(result, expected)
def test_estimate_spd_two_samples(self):
space = SPDMatrices(3)
metric = SPDMetricAffine(3)
point = space.random_point(2)
mean = FrechetMean(metric)
mean.fit(point)
result = mean.estimate_
expected = metric.exp(metric.log(point[0], point[1]) / 2, point[1])
self.assertAllClose(expected, result)
def __init__(self, n):
super(CorrelationMatricesBundle, self).__init__(
n=n,
ambient_metric=SPDMetricAffine(n),
group_dim=n,
group_action=FullRankCorrelationMatrices.diag_action,
)
def test_inverse_transform_spd(self):
point = SPDMatrices(3).random_uniform(10)
mean = FrechetMean(metric=SPDMetricAffine(3), point_type='matrix')
X = mean.fit_transform(X=point)
result = mean.inverse_transform(X)
expected = point
self.assertAllClose(expected, result)
def test_inverse_transform_spd(self):
point = SPDMatrices(3).random_uniform(10)
transformer = ToTangentSpace(geometry=SPDMetricAffine(3))
X = transformer.fit_transform(X=point)
result = transformer.inverse_transform(X)
expected = point
self.assertAllClose(expected, result, atol=1e-4)
def test_estimate_transform_spd(self):
point = SPDMatrices(3).random_uniform()
points = gs.stack([point, point])
transformer = ToTangentSpace(geometry=SPDMetricAffine(3))
transformer.fit(X=points)
result = transformer.transform(points)
expected = gs.zeros((2, 6))
self.assertAllClose(expected, result, atol=1e-5)
def test_estimate_spd(self):
point = SPDMatrices(3).random_point()
points = gs.array([point, point])
mean = FrechetMean(metric=SPDMetricAffine(3), point_type='matrix')
mean.fit(X=points)
result = mean.estimate_
expected = point
self.assertAllClose(expected, result)
def test_estimate_transform_spd(self):
point = SPDMatrices(3).random_uniform()
points = gs.array([point, point])
mean = FrechetMean(metric=SPDMetricAffine(3), point_type='matrix')
mean.fit(X=points)
result = mean.transform(points)
expected = gs.zeros((2, 6))
self.assertAllClose(expected, result)
def fit_test_data(self):
estimator = GeometricMedian(SPDMetricAffine(n=2))
X = gs.array([[[1.0, 0.0], [0.0, 1.0]], [[1.0, 0.0], [0.0, 1.0]]])
expected = gs.array([[1.0, 0.0], [0.0, 1.0]])
smoke_data = [dict(estimator=estimator, X=X, expected=expected)]
return self.generate_tests(smoke_data)
def setup_method(self):
"""Set up the test"""
self.n = 3
self.spd_cov_n = (self.n * (self.n + 1)) // 2
self.samples = 5
self.spd = SPDMatrices(self.n)
self.log_euclidean = SPDMetricLogEuclidean(self.n)
self.affine_invariant = SPDMetricAffine(self.n)
self.euclidean = Euclidean(self.n)
def setUp(self):
self.so3 = SpecialOrthogonal(n=3)
self.spd = SPDMatrices(3)
self.spd_metric = SPDMetricAffine(3)
self.n_samples = 10
self.X = self.so3.random_uniform(n_samples=self.n_samples)
self.metric = self.so3.bi_invariant_metric
self.n_components = 2
def setUp(self):
warnings.simplefilter('ignore', category=ImportWarning)
gs.random.seed(1234)
self.n = 3
self.space = SPDMatrices(n=self.n)
self.metric_affine = SPDMetricAffine(n=self.n)
self.metric_procrustes = SPDMetricProcrustes(n=self.n)
self.metric_euclidean = SPDMetricEuclidean(n=self.n)
self.metric_logeuclidean = SPDMetricLogEuclidean(n=self.n)
self.n_samples = 4
def setUp_alt(self, n=3, n_samples=4):
"""Set up the test, flexible parameters."""
warnings.simplefilter('ignore', category=ImportWarning)
gs.random.seed(1234)
self.n = n
self.space = SPDMatrices(n=self.n)
self.metric_affine = SPDMetricAffine(n=self.n)
self.metric_procrustes = SPDMetricProcrustes(n=self.n)
self.metric_euclidean = SPDMetricEuclidean(n=self.n)
self.metric_logeuclidean = SPDMetricLogEuclidean(n=self.n)
self.n_samples = n_samples
def setUp(self):
"""Set up the test."""
warnings.simplefilter('ignore', category=ImportWarning)
gs.random.seed(1234)
self.n = 3
self.space = SPDMatrices(n=self.n)
self.metric_affine = SPDMetricAffine(n=self.n)
self.metric_bureswasserstein = SPDMetricBuresWasserstein(n=self.n)
self.metric_euclidean = SPDMetricEuclidean(n=self.n)
self.metric_logeuclidean = SPDMetricLogEuclidean(n=self.n)
self.n_samples = 4
def main():
"""Execute illustration of MDM supervised classifier."""
n_samples = 100
n_features = 2
n_classes = 3
# generate toy dataset of 2D SPD matrices
dataset_generator = geomstats.datasets.sample_sdp_2d.DatasetSPD2D(
n_samples, n_features, n_classes)
data, labels = dataset_generator.generate_sample_dataset()
# plot dataset as ellipses
ellipsis = visualization.Ellipsis2D()
for i in range(n_samples):
x = data[i]
y = geomstats.datasets.sample_sdp_2d.get_label_at_index(i, labels)
ellipsis.draw(x, color=ellipsis.colors[y], alpha=.1)
# define and fit MDM classifier to data
metric = SPDMetricAffine(n=n_features)
MDMEstimator = RiemannianMinimumDistanceToMeanClassifier(
metric, n_classes, point_type='matrix')
MDMEstimator.fit(data, labels)
# plot Frechet means computed in the MDM
for i in range(n_classes):
ellipsis.draw(MDMEstimator.mean_estimates_[i],
color=ellipsis.colors_alt[i],
linewidth=5,
label='Barycenter of class ' + str(i))
# generate random test samples, and predict with MDM classifier
data_test = SPDMatrices(n=n_features).random_uniform(n_samples=3)
predictions = MDMEstimator.predict(data_test)
for i in range(data_test.shape[0]):
c = list(predictions[i] == 1).index(True)
x_from, y_from = ellipsis.draw(data_test[i],
color=ellipsis.colors[c],
linewidth=5)
_, _, x_to, y_to = ellipsis.compute_coordinates(
MDMEstimator.mean_estimates_[c])
arrow = visualization.DataArrow(ellipsis.fig)
arrow.draw(x_from, y_from, x_to, y_to)
ellipsis.fig.axes[0].set_title(
'Example plot of the MDM classifier in dimension 2\n'
'3-class fit and 3 test sample prediction\n'
'(black arrows denote assignement)')
ellipsis.plot()
def fit_sanity_test_data(self):
n = 4
estimator_1 = GeometricMedian(SPDMetricAffine(n))
space_1 = SPDMatrices(n)
space_2 = Hypersphere(2)
estimator_2 = GeometricMedian(space_2.metric)
smoke_data = [
dict(estimator=estimator_1, space=space_1),
dict(estimator=estimator_2, space=space_2),
]
return self.generate_tests(smoke_data)
def test_batched(self):
space = SPDMatrices(3)
metric = SPDMetricAffine(3)
point = space.random_point(4)
mean_batch = FrechetMean(metric, method="batch", verbose=True)
data = gs.stack([point[:2], point[2:]], axis=1)
mean_batch.fit(data)
result = mean_batch.estimate_
mean = FrechetMean(metric)
mean.fit(data[:, 0])
expected_1 = mean.estimate_
mean.fit(data[:, 1])
expected_2 = mean.estimate_
expected = gs.stack([expected_1, expected_2])
self.assertAllClose(expected, result)
def test_predict(self):
"""Test the predict method."""
n_clusters = 2
bary_a = gs.array([[EULER, 0], [0, 1]])
bary_b = gs.array([[EULER**4, 0], [0, 1]])
MDMEstimator = RiemannianMinimumDistanceToMeanClassifier(
SPDMetricAffine(n=2), n_clusters, point_type='matrix')
MDMEstimator.mean_estimates_ = gs.concatenate(
[bary_a[None, ...], bary_b[None, ...]])
X = gs.array([[EULER**3, 0], [0, 1]])[None, ...]
y_expected = gs.array([[0, 1]])
y_result = MDMEstimator.predict(X)
self.assertAllClose(y_result, y_expected)
def test_fit(self):
"""Test the fit method."""
n_clusters = 2
MDMEstimator = RiemannianMinimumDistanceToMeanClassifier(
SPDMetricAffine(n=2), n_clusters, point_type='matrix')
points_a = gs.array([[[EULER**2, 0], [0, 1]], [[1, 0], [0, 1]]])
labels_a = gs.array([[1, 0], [1, 0]])
bary_a_expected = gs.array([[EULER, 0], [0, 1]])
points_b = gs.array([[[EULER**8, 0], [0, 1]], [[1, 0], [0, 1]]])
labels_b = gs.array([[0, 1], [0, 1]])
bary_b_expected = gs.array([[EULER**4, 0], [0, 1]])
train_data = gs.concatenate([points_a, points_b])
train_labels = gs.concatenate([labels_a, labels_b])
MDMEstimator.fit(train_data, train_labels)
bary_a_result = MDMEstimator.mean_estimates_[0]
bary_b_result = MDMEstimator.mean_estimates_[1]
self.assertAllClose(bary_a_result, bary_a_expected)
self.assertAllClose(bary_b_result, bary_b_expected)
def test_log(self, n, power_affine, point, base_point, expected):
metric = SPDMetricAffine(n, power_affine)
self.assertAllClose(
metric.log(gs.array(point), gs.array(base_point)), gs.array(expected)
)
def test_exp(self, n, power_affine, tangent_vec, base_point, expected):
metric = SPDMetricAffine(n, power_affine)
self.assertAllClose(
metric.exp(gs.array(tangent_vec), gs.array(base_point)), gs.array(expected)
)
