def grade3():
marks = 0
try:
data = np.array([[i, i] for i in range(5)])
centers = np.array([[1., 1.], [2., 2.], [3., 3.]])
op = np.array([[0.5, 0.5], [2.0, 2.0], [3.5, 3.5]])
kmeans = KMeans(D=2, n_clusters=3)
kmeans.cluster_centers = centers
it = kmeans.train(data, 1)
if np.allclose(kmeans.cluster_centers, op) and it == 0:
marks += 0.5
data = np.array([[i + 1, i * 2.3] for i in range(5)])
centers = np.array([[5., 1.], [-1., 2.], [3., 6.]])
op = np.array([[5, 1], [1.5, 1.15], [4.0, 6.8999999999999995]])
kmeans = KMeans(D=2, n_clusters=3)
kmeans.cluster_centers = centers
it = kmeans.train(data, 1)
if np.allclose(kmeans.cluster_centers, op) and it == 0:
marks += 0.5
data = np.array([[i + 1, i * 2.3] for i in range(3)])
centers = np.array([[5, 1], [-1., 2]])
op = np.array([[3.0, 4.6], [1.5, 1.15]])
kmeans = KMeans(D=2, n_clusters=2)
kmeans.cluster_centers = centers
it = kmeans.train(data, 5)
if np.allclose(kmeans.cluster_centers, op) and it == 1:
marks += 1
except:
print('Error in k-means')
return marks
def test_update_centers(self):
"""
Tests update centers
"""
X, y, centers = generate_cluster_samples()
n_samples = X.shape[0]
n_features = X.shape[1]
k = centers.shape[0]
kmeans = KMeans(k, N_ITER)
# Set cluster centers so that assignment is deterministic
kmeans.cluster_centers = centers
assignments, distances = kmeans.assign_points(X)
assignments = kmeans.reinitialize_empty_clusters(
X, assignments, distances)
# clear out centers to test method
kmeans.cluster_centers = np.zeros((k, n_features))
kmeans.update_centers(X, assignments)
# calculate average difference in coordinates of estimated
# and real centers
error = np.linalg.norm(kmeans.cluster_centers - centers) / k
self.assertLess(error, EPS)
def test_assign_points(self):
"""
Tests initialize methods of the KMeans class.
"""
X, y, centers = generate_cluster_samples()
n_samples = X.shape[0]
k = centers.shape[0]
kmeans = KMeans(k, N_ITER)
# Set cluster centers so that assignment is deterministic
kmeans.cluster_centers = centers
assignments, distances = kmeans.assign_points(X)
# check assignment array shape
self.assertEqual(assignments.ndim, 1)
self.assertEqual(assignments.shape[0], n_samples)
# check distances array shape
self.assertEqual(distances.ndim, 1)
self.assertEqual(distances.shape[0], n_samples)
# check that assignments only include valid cluster indices (0 <= idx < k)
self.assertTrue(
np.all(np.logical_and(assignments < k, assignments >= 0)))
# Check cluster assignments are correct
self.assertTrue(np.all(assignments[:25] == 0))
self.assertTrue(np.all(assignments[25:50] == 1))
self.assertTrue(np.all(assignments[50:75] == 2))
self.assertTrue(np.all(assignments[75:] == 3))
def test_reinitialize_empty_clusters(self):
"""
Tests reassignment of points to empty clusters
"""
X, y, centers = generate_cluster_samples()
n_samples = X.shape[0]
k = centers.shape[0]
kmeans = KMeans(k, N_ITER)
# Set cluster centers so that assignment is deterministic
kmeans.cluster_centers = centers
assignments, distances = kmeans.assign_points(X)
# reassign all points in cluster 3 to cluster 2 to create empty cluster
assignments[75:] = 2
# reinitialize empty clusters by reassigning points
assignments = kmeans.reinitialize_empty_clusters(
X, assignments, distances)
# ensure that each cluster has an assigned point
# and that only valid cluster indices are used
self.assertSetEqual(set(assignments), set(range(k)))