def test_calc_shannon_entropy_length_error():
dataset = np.array([[1, 2, 3],
[4, 5, 6]])
labels = np.array([0, 0])
algo = ID3()
shannon_entropy = algo._calc_shannon_entropy(dataset, labels)
assert shannon_entropy == -1 * math.log(1)
def test_calc_shannon_entropy():
dataset = np.array([[1, 2, 3],
[4, 5, 6]])
labels = [0, 1]
algo = ID3()
shannon_entropy = algo._calc_shannon_entropy(dataset, labels)
assert shannon_entropy == -0.5 * math.log(1 / 2) * 2
def test_create_tree():
algo = ID3()
dataset = np.array([[1, 1, 0],
[0, 1, 0],
[1, 1, 1]])
labels = np.array([0, 0, 1])
tree = algo._create_tree(dataset, labels)
def test_choose_best_feature():
algo = ID3()
dataset = np.array([[1, 1, 0],
[0, 1, 0],
[1, 1, 1]])
labels = np.array([0, 0, 1])
best_feature = algo._choose_best_feature(dataset, labels)
assert best_feature == 2
def test_calc_shannon_entropy_03(capsys):
dataset = np.array([[1, 2, 3],
[4, 5, 6]])
labels = np.array([0])
algo = ID3()
try:
shannon_entropy = algo._calc_shannon_entropy(dataset, labels)
except Exception as e:
captured = capsys.readouterr()
assert captured.out == "dataset length doesn't equal with labels\n"
print(e)
def test_conditional_entropy():
dataset = np.array([[1, 2, 3],
[4, 5, 6]])
labels = np.array([0, 1])
algo = ID3()
conditional_entropy = algo._calc_conditional_entropy(dataset, labels, 0)
assert conditional_entropy == 0
dataset = np.array([[1, 1, 0],
[0, 1, 0],
[1, 1, 1]])
labels = np.array([0, 0, 1])
conditional_entropy = algo._calc_conditional_entropy(dataset, labels, 0)
target_entropy = 1 / 3 * (-1 * math.log(1)) + 2 / 3 * (-1 / 2 * math.log(1 / 2) * 2)
assert math.isclose(conditional_entropy, target_entropy)
def test_find_max_class():
algo = ID3()
labels = np.array([0])
max_label = algo._find_max_class(labels)
assert max_label == 0
labels = np.array([0, 0, 1])
max_label = algo._find_max_class(labels)
assert max_label == 0
labels = np.array([0, 1, 1])
max_label = algo._find_max_class(labels)
assert max_label == 1
labels = np.array([0, 1, 1, 2, 2, 2])
max_label = algo._find_max_class(labels)
assert max_label == 2
def test_split_dateset():
dataset = np.array([[1, 1, 0, 1],
[0, 0, 1, 1],
[1, 1, 1, 0]])
algo = ID3()
new_dataset = algo._split_dateset(dataset, 0, 0)
assert (new_dataset == np.array([[0, 1, 1]])).all()
new_dataset = algo._split_dateset(dataset, 0, 1)
assert (new_dataset == np.array([[1, 0, 1],
[1, 1, 0]])).all()
new_dataset = algo._split_dateset(dataset, 2, 1)
assert (new_dataset == np.array([[0, 0, 1],
[1, 1, 0]])).all()
new_dataset = algo._split_dateset(dataset, 3, 0)
assert (new_dataset == np.array([[1, 1, 1]])).all()
def test_split_label():
dataset = np.array([[1, 1, 0, 1],
[0, 0, 1, 1],
[1, 1, 1, 0],
[1, 1, 0, 1],
[0, 0, 1, 1],
[1, 1, 1, 0]])
labels = np.array([1, 1, 1, 0, 0, 1])
algo = ID3()
new_label = algo._split_labels(dataset, labels, 1, 1)
assert (new_label == np.array([1, 1, 0, 1])).all()
new_label = algo._split_labels(dataset, labels, 1, 0)
assert (new_label == np.array([1, 0])).all()
new_label = algo._split_labels(dataset, labels, 3, 1)
assert (new_label == np.array([1, 1, 0, 0])).all()