def test_feature_prob_exception_not_sum_one(self):
with self.assertRaises(ValueError):
self.decision_tree = DecisionTreeClassifier(
feature_prob=[0.2, 0.4, 0.5])
def test_feature_prob_none_value(self):
self.decision_tree = DecisionTreeClassifier(feature_prob=None)
self.assertIsNone(self.decision_tree.feature_prob)
def test_feature_prob_exception_negative_values(self):
with self.assertRaises(ValueError):
self.decision_tree = DecisionTreeClassifier(
feature_prob=[-0.2, 0.4, 0.4])
def test_feature_selection_exception_non_admissible_value(self):
with self.assertRaises(ValueError):
self.decision_tree = DecisionTreeClassifier(
feature_selection='non')
def test_min_samples_split_positive_value_greater_than_one(self):
self.decision_tree = DecisionTreeClassifier(min_samples_split=2)
self.assertEqual(self.decision_tree.min_samples_split, 2)
def test_min_gain_split_non_negative_value(self):
self.decision_tree = DecisionTreeClassifier(min_gain_split=1)
self.assertEqual(self.decision_tree.min_gain_split, 1)
def test_split_chooser_admissible_value(self):
self.decision_tree = DecisionTreeClassifier(split_chooser='best')
self.assertIsInstance(self.decision_tree._split_chooser,
BestSplitChooser)
def test_max_depth_positive_value(self):
self.decision_tree = DecisionTreeClassifier(max_depth=1)
self.assertEqual(self.decision_tree.max_depth, 1)
def test_min_samples_leaf_exception_negative_value(self):
with self.assertRaises(ValueError):
self.decision_tree = DecisionTreeClassifier(min_samples_leaf=-1)
def test_max_depth_exception_negative_value(self):
with self.assertRaises(ValueError):
self.decision_tree = DecisionTreeClassifier(max_depth=-1)
def test_max_depth_none_value(self):
self.decision_tree = DecisionTreeClassifier(max_depth=None)
self.assertIsNone(self.decision_tree.max_depth)
def setUp(self):
self.decision_tree = DecisionTreeClassifier()
self.X = np.array(['A', 'B', 'A', 'B', 'B', 'C', 'A', 'C',
'B']).reshape((3, 3))
self.y = np.array([1, 1, 0])
class DecisionTreeClassifierTest(TestCase):
def setUp(self):
self.decision_tree = DecisionTreeClassifier()
self.X = np.array(['A', 'B', 'A', 'B', 'B', 'C', 'A', 'C',
'B']).reshape((3, 3))
self.y = np.array([1, 1, 0])
def tearDown(self):
pass
def test_fit(self):
self.decision_tree.fit(self.X, self.y)
self.assertIsNotNone(self.decision_tree._encoder)
self.assertIsNotNone(self.decision_tree._tree_builder)
self.assertIsNotNone(self.decision_tree._tree)
self.assertIsInstance(self.decision_tree._tree, DecisionTree)
print(self.decision_tree._tree.nodes)
def test_predict_empty_vector(self):
self.decision_tree.fit(self.X, self.y)
x = np.array([]).reshape(0, 0)
with self.assertRaises(ValueError):
self.decision_tree.predict(x)
def test_predict_one_instance(self):
self.decision_tree._tree = mock.MagicMock(spec=DecisionTree)
self.decision_tree._tree.predict.return_value = 1
self.decision_tree._n_features = 3
self.decision_tree._encoder = mock.MagicMock(spec=LabelEncoder)
self.decision_tree._encoder.inverse_transform.return_value = [1]
x = np.array(['A', 'B', 'A']).reshape((1, 3))
expected_length_prediction = 1
resulted_length_prediction = len(self.decision_tree.predict(x))
self.assertEqual(resulted_length_prediction,
expected_length_prediction)
def test_predict_two_instances(self):
self.decision_tree._tree = mock.MagicMock(spec=DecisionTree)
self.decision_tree._tree.predict.side_effect = [1, 0]
self.decision_tree._n_features = 3
self.decision_tree._encoder = mock.MagicMock(spec=LabelEncoder)
self.decision_tree._encoder.inverse_transform.return_value = [1, 0]
x = np.array(['A', 'B', 'A', 'C', 'C', 'A']).reshape((2, 3))
expected_prediction = 2
resulted_prediction = len(self.decision_tree.predict(x))
self.assertEqual(resulted_prediction, expected_prediction)
def test_predict_proba_one_instance(self):
self.decision_tree._tree = mock.MagicMock(spec=DecisionTree)
self.decision_tree._tree.predict_proba.return_value = [0.25, 0.75]
self.decision_tree._n_features = 3
x = np.array(['A', 'B', 'A']).reshape((1, 3))
expected_length_prediction = 2
resulted_prediction = self.decision_tree.predict_proba(x)
self.assertEqual(len(resulted_prediction), 1)
for resulted in resulted_prediction:
self.assertEqual(len(resulted), expected_length_prediction)
def test_predict_proba_two_instance(self):
self.decision_tree._tree = mock.MagicMock(spec=DecisionTree)
self.decision_tree._tree.predict_proba.return_value = [[0.25, 0.75],
[0.33, 0.67]]
self.decision_tree._n_features = 3
x = np.array(['A', 'B', 'A', 'C', 'C', 'A']).reshape((2, 3))
expected_length_prediction = 2
resulted_prediction = self.decision_tree.predict_proba(x)
self.assertEqual(len(resulted_prediction), 2)
for resulted in resulted_prediction:
self.assertEqual(len(resulted), expected_length_prediction)
def test_feature_selection_admissible_value(self):
self.decision_tree = DecisionTreeClassifier(feature_selection='all')
self.assertIsInstance(self.decision_tree._feature_selection,
AllFeatureSelection)
def test_feature_prob_positive_values_sum_one(self):
self.decision_tree = DecisionTreeClassifier(
feature_prob=[0.25, 0.25, 0.25, 0.25])
self.assertEqual(self.decision_tree.feature_prob,
[0.25, 0.25, 0.25, 0.25])
def test_min_samples_leaf_positive_value(self):
self.decision_tree = DecisionTreeClassifier(min_samples_leaf=1)
self.assertEqual(self.decision_tree.min_samples_leaf, 1)
def test_min_gain_split_exception_negative_value(self):
with self.assertRaises(ValueError):
self.decision_tree = DecisionTreeClassifier(min_gain_split=-1)
def test_min_samples_split_exception_none_value(self):
with self.assertRaises(ValueError):
self.decision_tree = DecisionTreeClassifier(min_samples_split=None)
def test_split_chooser_exception_non_admissible_value(self):
with self.assertRaises(ValueError):
self.decision_tree = DecisionTreeClassifier(split_chooser='non')
def test_min_samples_split_exception_less_than_two_instances(self):
with self.assertRaises(ValueError):
self.decision_tree = DecisionTreeClassifier(min_samples_split=1)
def test_split_criterion_exception_none_value(self):
with self.assertRaises(ValueError):
self.decision_tree = DecisionTreeClassifier(split_criterion=None)
def test_split_criterion_admissible_value(self):
self.decision_tree = DecisionTreeClassifier(split_criterion='gini')
self.assertIsInstance(self.decision_tree._split_criterion,
GiniCriterion)