def test_label_mrr_zeros():
"""Test MRR where no relevant labels"""
y_true = np.array([[0, 0, 0], [0, 0, 0], [0, 0, 0]])
y_prob = np.array([[0.3, 0.7, 0.0], [0.1, 0.0, 0.0], [0.1, 0.5, 0.3]])
# Check results
result = metriks.label_mean_reciprocal_rank(y_true, y_prob)
np.testing.assert_equal(result.mask.all(), True)
print(result)
def test_label_mrr_with_errors():
y_true = np.array([[0, 1, 1, 0], [1, 1, 1, 0], [0, 1, 1, 1]])
y_prob = np.array([[0.2, 0.9, 0.8, 0.4], [0.9, 0.2, 0.8, 0.4],
[0.2, 0.8, 0.9, 0.4]])
result = metriks.label_mean_reciprocal_rank(y_true, y_prob)
expected = np.ma.array([1.0, (7 / 4) / 3, 2 / 3, 1 / 3])
np.testing.assert_allclose(result, expected)
print(result)
def test_label_mrr_ones():
"""Test MRR where all labels are relevant and predictions are perfect"""
y_true = np.array([[1, 1, 1], [1, 1, 1], [1, 1, 1]])
y_prob = np.array([[0.3, 0.7, 0.0], [0.1, 0.9, 0.0], [0.2, 0.5, 0.1]])
# Check results
expected = np.ma.array([0.5, 1.0, 1 / 3])
result = metriks.label_mean_reciprocal_rank(y_true, y_prob)
np.testing.assert_allclose(result, expected)
print(result)
def test_label_mrr_some_zeros():
"""Test MRR where some relevant labels"""
y_true = np.array([[0, 1, 0], [0, 0, 0], [0, 1, 0]])
y_prob = np.array([[0.3, 0.7, 0.0], [0.1, 0.0, 0.0], [0.1, 0.5, 0.3]])
# Check results
expected = np.ma.array([0.0, 1.0, 0.0], mask=[True, False, True])
result = metriks.label_mean_reciprocal_rank(y_true, y_prob)
np.testing.assert_allclose(result, expected)
print(result)
def test_label_mrr_perfect():
"""Test MRR where the probabilities are perfect"""
y_true = np.array([
[0, 1, 0],
[1, 0, 0],
[0, 1, 0],
[1, 0, 0],
[0, 0, 1],
])
y_prob = np.array([
[0.3, 0.7, 0.0],
[0.1, 0.0, 0.0],
[0.1, 0.5, 0.3],
[0.6, 0.2, 0.4],
[0.1, 0.2, 0.3],
])
# Check results
expected = np.ma.array([1.0, 0.0, 0.0], mask=[False, True, True])
result = metriks.label_mean_reciprocal_rank(y_true, y_prob)
np.testing.assert_allclose(result, expected)
print(result)