def test_hypersphere_kmeans_init_array(self):
n_features = 4
n_clusters = 3
manifold = hypersphere.Hypersphere(n_features - 1)
centroids = manifold.random_von_mises_fisher(kappa=10, n_samples=n_clusters)
self._test_hypersphere_kmeans_init(
centroids, n_features=n_features, n_clusters=n_clusters
)
def test_hypersphere_kmedoids_fit(self):
gs.random.seed(55)
manifold = hypersphere.Hypersphere(2)
metric = hypersphere.HypersphereMetric(2)
data = manifold.random_von_mises_fisher(kappa=100, n_samples=200)
kmedoids = RiemannianKMedoids(metric=metric, n_clusters=1)
center = kmedoids.fit(data)
self.assertTrue(manifold.belongs(center))
def test_hypersphere_kmeans_fit(self):
gs.random.seed(55)
manifold = hypersphere.Hypersphere(2)
metric = hypersphere.HypersphereMetric(2)
x = manifold.random_von_mises_fisher(kappa=100, n_samples=200)
kmeans = RiemannianKMeans(metric, 1, tol=1e-3)
kmeans.fit(x)
center = kmeans.centroids
mean = metric.mean(x)
result = metric.dist(center, mean)
expected = 0.
self.assertAllClose(expected, result, atol=1e-2)
def test_hypersphere_kmeans_predict(self):
gs.random.seed(1234)
manifold = hypersphere.Hypersphere(2)
metric = hypersphere.HypersphereMetric(2)
x = manifold.random_von_mises_fisher(kappa=100, n_samples=200)
kmeans = RiemannianKMeans(metric, 5, tol=1e-5)
kmeans.fit(x, max_iter=100)
result = kmeans.predict(x)
centroids = kmeans.centroids
expected = gs.array([int(metric.closest_neighbor_index(x_i, centroids))
for x_i in x])
self.assertAllClose(expected, result)
def test_hypersphere_kmedoids_predict(self):
gs.random.seed(1234)
dim = 2
manifold = hypersphere.Hypersphere(dim)
metric = hypersphere.HypersphereMetric(dim)
data = manifold.random_von_mises_fisher(kappa=100, n_samples=200)
kmedoids = RiemannianKMedoids(metric, n_clusters=5)
centroids = kmedoids.fit(data, max_iter=100)
result = kmedoids.predict(data)
expected = gs.array([
int(metric.closest_neighbor_index(x_i, centroids)) for x_i in data
])
self.assertAllClose(expected, result)
def test_hypersphere_kmeans_initialization(self):
gs.random.seed(55)
manifold = hypersphere.Hypersphere(2)
metric = hypersphere.HypersphereMetric(2)
x = manifold.random_von_mises_fisher(kappa=100, n_samples=200)
n_clusters = 3
kmeans = RiemannianKMeans(
metric, n_clusters, init_step_size=1.0, tol=1e-3, init="kmeans++"
)
kmeans.fit(x)
centroids = kmeans.centroids
result = centroids.shape
expected = (n_clusters, 3)
self.assertAllClose(expected, result)
def test_hypersphere_kmeans_fit(self):
gs.random.seed(55)
manifold = hypersphere.Hypersphere(2)
metric = hypersphere.HypersphereMetric(2)
x = manifold.random_von_mises_fisher(kappa=100, n_samples=200)
kmeans = RiemannianKMeans(metric, 1, init_step_size=1.0, tol=1e-3)
kmeans.fit(x)
center = kmeans.centroids
mean = FrechetMean(metric=metric, init_step_size=1.0)
mean.fit(x)
result = metric.dist(center, mean.estimate_)
expected = 0.0
self.assertAllClose(expected, result)
def _test_hypersphere_kmeans_init(
self, init, *, n_features=4, n_clusters=3, seed=1
):
gs.random.seed(seed)
manifold = hypersphere.Hypersphere(n_features - 1)
x = manifold.random_von_mises_fisher(kappa=100, n_samples=200)
kmeans = RiemannianKMeans(
manifold.metric, n_clusters, init_step_size=1.0, tol=1e-3, init=init
)
kmeans.fit(x)
centroids = kmeans.centroids
result = centroids.shape
expected = (n_clusters, n_features)
self.assertAllClose(expected, result)
def make_centroids(X, n_clusters):
n_features = X.shape[1]
manifold = hypersphere.Hypersphere(n_features - 1)
centroids = manifold.random_von_mises_fisher(kappa=10, n_samples=n_clusters)
return centroids
