def test_locate_elbow(self):
"""
Test the addition of locate_elbow to an image
"""
X, y = make_blobs(n_samples=1000,
n_features=5,
centers=3,
shuffle=True,
random_state=42)
visualizer = KElbowVisualizer(
KMeans(random_state=0),
k=6,
metric="calinski_harabasz",
timings=False,
locate_elbow=True,
)
visualizer.fit(X)
assert len(visualizer.k_scores_) == 5
assert visualizer.elbow_value_ == 3
expected = np.array(
[4286.479848, 12463.383743, 8766.999551, 6950.08391, 5865.79722])
visualizer.finalize()
self.assert_images_similar(visualizer, windows_tol=2.2)
assert_array_almost_equal(visualizer.k_scores_, expected)
def test_timings(self):
"""
Test the twinx double axes with k-elbow timings
"""
visualizer = KElbowVisualizer(KMeans(random_state=0),
k=5,
timings=True,
locate_elbow=False)
visualizer.fit(self.clusters.X)
# Check that we kept track of time
assert len(visualizer.k_timers_) == 4
assert all([t > 0 for t in visualizer.k_timers_])
# Check that we plotted time on a twinx
assert hasattr(visualizer, "axes")
assert len(visualizer.axes) == 2
# delete the timings axes and
# overwrite k_timers_, k_values_ for image similarity Tests
visualizer.axes[1].remove()
visualizer.k_timers_ = [
0.01084589958190918,
0.011144161224365234,
0.017028093338012695,
0.010634183883666992,
]
visualizer.k_values_ = [2, 3, 4, 5]
# call draw again which is normally called in fit
visualizer.draw()
visualizer.finalize()
self.assert_images_similar(visualizer)
def test_integrated_mini_batch_kmeans_elbow(self):
"""
Test no exceptions for mini-batch kmeans k-elbow visualizer
"""
# NOTE #182: cannot use occupancy dataset because of memory usage
# Generate a blobs data set
X, y = make_blobs(n_samples=1000,
n_features=12,
centers=6,
shuffle=True,
random_state=42)
try:
_, ax = plt.subplots()
visualizer = KElbowVisualizer(MiniBatchKMeans(random_state=42),
k=4,
ax=ax)
visualizer.fit(X)
visualizer.finalize()
self.assert_images_similar(visualizer)
except Exception as e:
pytest.fail("error during k-elbow: {}".format(e))
def test_calinski_harabasz_metric(self):
"""
Test the calinski-harabasz metric of the k-elbow visualizer
"""
visualizer = KElbowVisualizer(
KMeans(random_state=0),
k=5,
metric="calinski_harabasz",
timings=False,
locate_elbow=False,
)
visualizer.fit(self.clusters.X)
assert len(visualizer.k_scores_) == 4
assert visualizer.elbow_value_ is None
expected = np.array([
81.66272625603568,
50.992378259195554,
39.573201061900455,
37.06865804955547,
])
visualizer.finalize()
self.assert_images_similar(visualizer)
assert_array_almost_equal(visualizer.k_scores_, expected)
def test_topic_modeling_k_means(self):
"""
Test topic modeling k-means on the hobbies corpus
"""
corpus = load_hobbies()
tfidf = TfidfVectorizer()
docs = tfidf.fit_transform(corpus.data)
visualizer = KElbowVisualizer(KMeans(), k=(4, 8))
visualizer.fit(docs)
visualizer.finalize()
self.assert_images_similar(visualizer)
def test_silhouette_metric(self):
"""
Test the silhouette metric of the k-elbow visualizer
"""
visualizer = KElbowVisualizer(
KMeans(random_state=0),
k=5,
metric="silhouette",
timings=False,
locate_elbow=False,
)
visualizer.fit(self.clusters.X)
expected = np.array([0.691636, 0.456646, 0.255174, 0.239842])
assert len(visualizer.k_scores_) == 4
visualizer.finalize()
self.assert_images_similar(visualizer)
assert_array_almost_equal(visualizer.k_scores_, expected)
def test_distortion_metric(self):
"""
Test the distortion metric of the k-elbow visualizer
"""
visualizer = KElbowVisualizer(
KMeans(random_state=0),
k=5,
metric="distortion",
timings=False,
locate_elbow=False,
)
visualizer.fit(self.clusters.X)
expected = np.array([69.100065, 54.081571, 43.146921, 34.978487])
assert len(visualizer.k_scores_) == 4
visualizer.finalize()
self.assert_images_similar(visualizer, tol=0.03)
assert_array_almost_equal(visualizer.k_scores_, expected)
def test_calinski_harabasz_metric(self):
"""
Test the calinski-harabasz metric of the k-elbow visualizer
"""
visualizer = KElbowVisualizer(
KMeans(random_state=0),
k=5,
metric="calinski_harabasz",
timings=False,
locate_elbow=False,
)
visualizer.fit(self.clusters.X)
assert len(visualizer.k_scores_) == 4
assert visualizer.elbow_value_ is None
expected = np.array([81.662726, 50.992378, 40.952179, 35.939494])
visualizer.finalize()
self.assert_images_similar(visualizer)
assert_array_almost_equal(visualizer.k_scores_, expected)
def test_silhouette_metric(self):
"""
Test the silhouette metric of the k-elbow visualizer
"""
visualizer = KElbowVisualizer(
KMeans(random_state=0),
k=5,
metric="silhouette",
timings=False,
locate_elbow=False,
)
visualizer.fit(self.clusters.X)
expected = np.array([
0.6916363804000003,
0.456645663683503,
0.26918583373704463,
0.25523298106687914,
])
assert len(visualizer.k_scores_) == 4
visualizer.finalize()
self.assert_images_similar(visualizer)
assert_array_almost_equal(visualizer.k_scores_, expected)
def test_distortion_metric(self):
"""
Test the distortion metric of the k-elbow visualizer
"""
visualizer = KElbowVisualizer(
KMeans(random_state=0),
k=5,
metric="distortion",
timings=False,
locate_elbow=False,
)
visualizer.fit(self.clusters.X)
expected = np.array([
69.10006514142941,
54.081571290449936,
44.491830981793605,
33.99887993254433,
])
assert len(visualizer.k_scores_) == 4
visualizer.finalize()
self.assert_images_similar(visualizer, tol=0.03)
assert_array_almost_equal(visualizer.k_scores_, expected)
def test_set_colors_manually(self):
"""
Test the silhouette metric of the k-elbow visualizer
"""
oz = KElbowVisualizer(
KMeans(random_state=0),
k=5,
)
oz.metric_color = "r"
oz.timing_color = "y"
oz.vline_color = "c"
# Create artificial "fit" data for testing purposes
oz.k_values_ = [1, 2, 3, 4, 5, 6, 7, 8]
oz.k_timers_ = [6.2, 8.3, 10.1, 15.8, 21.2, 27.9, 38.2, 44.9]
oz.k_scores_ = [.8, .7, .55, .48, .40, .38, .35, .30]
oz.elbow_value_ = 5
oz.elbow_score_ = 0.40
# Execute drawing
oz.draw()
oz.finalize()
self.assert_images_similar(oz, tol=3.2)
