def test_nans_filled_with_zero():
"""
Check that nans are filled with zero
"""
arr = np.array([np.nan, 1., 2., 3.])
assert (CHAID.ContinuousColumn(arr).arr == np.array([0., 1., 2.,
3.])).all()
def test_chaid_vector_with_dtype_object():
"""
Check that error raised when object column supplied
"""
with pytest.raises(ValueError) as excinfo:
vector = CHAID.ContinuousColumn(np.array(['2', '4'], dtype="object"))
assert excinfo.value.message == 'Must only pass numerical values to create continuous column'
def test_chaid_vector_converts_strings():
"""
Check that error raised when string column supplied
"""
with pytest.raises(ValueError) as excinfo:
vector = CHAID.ContinuousColumn(np.array(['2', '4']))
assert excinfo.value.message == 'Must only pass numerical values to create continuous column'
def test_dependent_variable_properties_as_members_for_continous_node():
"""
Tests that node prints the std and mean when contonuous column supplied
"""
data = np.array([1.3, 23.0, 3.0, 3.0, 12.4])
continuous_dp = CHAID.ContinuousColumn(data)
node = CHAID.Node(dep_v=continuous_dp)
assert node.members == {'mean': data.mean(), 's.t.d': data.std()}
def test_node_should_have_a_score():
"""
Tests that node uses the split score and is called 'score'
"""
data = np.array([1.3, 23.0, 3.0, 3.0, 12.4])
continuous_dp = CHAID.ContinuousColumn(data)
split = CHAID.Split("a", [], 2, 3, 4)
node = CHAID.Node(dep_v=continuous_dp, split=split)
assert node.score == 2
def test_p_and_chi_values_for_normal_data(self):
"""
Check chi and p value against hand calculated values
"""
split = self.stats_normal_data.best_con_split(
self.ndarr,
CHAID.ContinuousColumn(self.normal_arr)
)
assert round(split.score, 4) == 2.7346
assert round(split.p, 4) == 0.0982
assert split.dof == 118.
def test_p_and_chi_values_for_random_data(self):
"""
Check chi and p value against hand calculated values
"""
split = self.stats_random_data.best_con_split(
self.ndarr,
CHAID.ContinuousColumn(self.random_arr)
)
assert round(split.score, 4) == 1.0588
assert round(split.p, 4) == 0.3056
assert split.dof == 118.
def test_column_stores_weights():
"""
Tests that the columns store the weights when they are passed
"""
arr = np.array([1.0, 2.0, 3.0])
wt = np.array([2.0, 1.0, 0.25])
nominal = CHAID.NominalColumn(arr, weights=wt)
ordinal = CHAID.OrdinalColumn(arr, weights=wt)
continuous = CHAID.ContinuousColumn(arr, weights=wt)
assert (nominal.weights == wt).all()
assert (ordinal.weights == wt).all()
assert (continuous.weights == wt).all()