def test_get_distances_to_centroids(self):
dataset = DataSet([[1, 5], [2, 1], [6, 5]])
centroids = [pd.Series([4, 5]), pd.Series([6, 2])]
results = clustering._get_distances_to_centroids(
dataset, centroids, euclidean)
assert_that(
results,
equals_dataframe([[3, 5.83], [4.47, 4.12], [2, 3]], places=2))
def test_get_distances_to_centroids(self):
dataset = DataSet([[1, 5], [2, 1], [6, 5]])
centroids = [pd.Series([4, 5]), pd.Series([6, 2])]
results = clustering._get_distances_to_centroids(dataset, centroids,
euclidean)
assert_that(results,
equals_dataframe([[3, 5.83], [4.47, 4.12], [2, 3]],
places=2))
def test_get_weights(self):
dataset = self.create_otago_dataset()
reduced = pca.pca(dataset, 2)
# Note: this is mostly a regression test, these expected values were
# obtained by running the code.
assert_that(
pd.DataFrame(reduced.get_weights()),
equals_dataframe(
[[-0.6778734, -0.73517866], [-0.73517866, 0.6778734]],
places=3))
def test_get_data_frame(self):
dataset = DataSet([[1, 2], [3, 4], [5, 6], [7, 8]],
labels=pd.Series(["b", "b", "b", "a"]))
df = dataset.get_data_frame()
assert_that(df, equals_dataframe([[1, 2], [3, 4], [5, 6], [7, 8]]))
def test_get_labelled_data_frame_no_labels(self):
dataset = DataSet([[1, 2], [3, 4], [5, 6], [7, 8]])
df = dataset.get_labelled_data_frame()
assert_that(df, equals_dataframe([[1, 2], [3, 4], [5, 6], [7, 8]]))
