def test_mc_t_two_sample(self):
"""Test gives correct results with valid input data."""
# Verified against R's t.test() and Deducer::perm.t.test().
# With numpy array as input.
exp = (-0.11858541225631833, 0.90756579317867436)
I = np.array([7.2, 7.1, 9.1, 7.2, 7.3, 7.2, 7.5])
II = np.array([8.8, 7.5, 7.7, 7.6, 7.4, 6.7, 7.2])
obs = mc_t_two_sample(I, II)
np.testing.assert_allclose(obs[:2], exp)
self.assertEqual(len(obs[2]), 999)
self.assertCorrectPValue(0.8, 0.9, mc_t_two_sample, [I, II],
p_val_idx=3)
# With python list as input.
exp = (-0.11858541225631833, 0.90756579317867436)
I = [7.2, 7.1, 9.1, 7.2, 7.3, 7.2, 7.5]
II = [8.8, 7.5, 7.7, 7.6, 7.4, 6.7, 7.2]
obs = mc_t_two_sample(I, II)
np.testing.assert_allclose(obs[:2], exp)
self.assertEqual(len(obs[2]), 999)
self.assertCorrectPValue(0.8, 0.9, mc_t_two_sample, [I, II],
p_val_idx=3)
exp = (-0.11858541225631833, 0.45378289658933718)
obs = mc_t_two_sample(I, II, tails='low')
np.testing.assert_allclose(obs[:2], exp)
self.assertEqual(len(obs[2]), 999)
self.assertCorrectPValue(0.4, 0.47, mc_t_two_sample, [I, II],
{'tails': 'low'}, p_val_idx=3)
exp = (-0.11858541225631833, 0.54621710341066287)
obs = mc_t_two_sample(I, II, tails='high', permutations=99)
np.testing.assert_allclose(obs[:2], exp)
self.assertEqual(len(obs[2]), 99)
self.assertCorrectPValue(0.4, 0.62, mc_t_two_sample, [I, II],
{'tails': 'high', 'permutations': 99},
p_val_idx=3)
exp = (-2.8855783649036986, 0.99315596652421401)
obs = mc_t_two_sample(I, II, tails='high',
permutations=99, exp_diff=1)
np.testing.assert_allclose(obs[:2], exp)
self.assertEqual(len(obs[2]), 99)
self.assertCorrectPValue(0.55, 0.99, mc_t_two_sample, [I, II],
{'tails': 'high', 'permutations': 99,
'exp_diff': 1}, p_val_idx=3)
def test_mc_t_two_sample(self):
"""Test gives correct results with valid input data."""
# Verified against R's t.test() and Deducer::perm.t.test().
# With numpy array as input.
exp = (-0.11858541225631833, 0.90756579317867436)
I = np.array([7.2, 7.1, 9.1, 7.2, 7.3, 7.2, 7.5])
II = np.array([8.8, 7.5, 7.7, 7.6, 7.4, 6.7, 7.2])
obs = mc_t_two_sample(I, II)
np.testing.assert_allclose(obs[:2], exp)
self.assertEqual(len(obs[2]), 999)
self.assertCorrectPValue(0.8, 0.9, mc_t_two_sample, [I, II],
p_val_idx=3)
# With python list as input.
exp = (-0.11858541225631833, 0.90756579317867436)
I = [7.2, 7.1, 9.1, 7.2, 7.3, 7.2, 7.5]
II = [8.8, 7.5, 7.7, 7.6, 7.4, 6.7, 7.2]
obs = mc_t_two_sample(I, II)
np.testing.assert_allclose(obs[:2], exp)
self.assertEqual(len(obs[2]), 999)
self.assertCorrectPValue(0.8, 0.9, mc_t_two_sample, [I, II],
p_val_idx=3)
exp = (-0.11858541225631833, 0.45378289658933718)
obs = mc_t_two_sample(I, II, tails='low')
np.testing.assert_allclose(obs[:2], exp)
self.assertEqual(len(obs[2]), 999)
self.assertCorrectPValue(0.4, 0.47, mc_t_two_sample, [I, II],
{'tails': 'low'}, p_val_idx=3)
exp = (-0.11858541225631833, 0.54621710341066287)
obs = mc_t_two_sample(I, II, tails='high', permutations=99)
np.testing.assert_allclose(obs[:2], exp)
self.assertEqual(len(obs[2]), 99)
self.assertCorrectPValue(0.4, 0.62, mc_t_two_sample, [I, II],
{'tails': 'high', 'permutations': 99},
p_val_idx=3)
exp = (-2.8855783649036986, 0.99315596652421401)
obs = mc_t_two_sample(I, II, tails='high',
permutations=99, exp_diff=1)
np.testing.assert_allclose(obs[:2], exp)
self.assertEqual(len(obs[2]), 99)
self.assertCorrectPValue(0.55, 0.99, mc_t_two_sample, [I, II],
{'tails': 'high', 'permutations': 99,
'exp_diff': 1}, p_val_idx=3)
def test_mc_t_two_sample_no_perms(self):
"""Test gives empty permutation results if no perms are given."""
exp = (-0.11858541225631833, 0.90756579317867436, [], None)
I = np.array([7.2, 7.1, 9.1, 7.2, 7.3, 7.2, 7.5])
II = np.array([8.8, 7.5, 7.7, 7.6, 7.4, 6.7, 7.2])
obs = mc_t_two_sample(I, II, permutations=0)
np.testing.assert_allclose(obs[0], exp[0])
np.testing.assert_allclose(obs[1], exp[1])
self.assertEqual(obs[2], exp[2])
self.assertEqual(obs[3], exp[3])
def test_mc_t_two_sample_no_perms(self):
"""Test gives empty permutation results if no perms are given."""
exp = (-0.11858541225631833, 0.90756579317867436, [], None)
I = np.array([7.2, 7.1, 9.1, 7.2, 7.3, 7.2, 7.5])
II = np.array([8.8, 7.5, 7.7, 7.6, 7.4, 6.7, 7.2])
obs = mc_t_two_sample(I, II, permutations=0)
np.testing.assert_allclose(obs[0], exp[0])
np.testing.assert_allclose(obs[1], exp[1])
self.assertEqual(obs[2], exp[2])
self.assertEqual(obs[3], exp[3])
def test_mc_t_two_sample_unbalanced_obs(self):
"""Test gives correct results with unequal number of obs per sample."""
# Verified against R's t.test() and Deducer::perm.t.test().
exp = (-0.10302479888889175, 0.91979753020527177)
I = np.array([7.2, 7.1, 9.1, 7.2, 7.3, 7.2])
II = np.array([8.8, 7.5, 7.7, 7.6, 7.4, 6.7, 7.2])
obs = mc_t_two_sample(I, II)
np.testing.assert_allclose(obs[:2], exp)
self.assertEqual(len(obs[2]), 999)
self.assertCorrectPValue(0.8, 0.9, mc_t_two_sample, [I, II],
p_val_idx=3)
def test_mc_t_two_sample_unbalanced_obs(self):
"""Test gives correct results with unequal number of obs per sample."""
# Verified against R's t.test() and Deducer::perm.t.test().
exp = (-0.10302479888889175, 0.91979753020527177)
I = np.array([7.2, 7.1, 9.1, 7.2, 7.3, 7.2])
II = np.array([8.8, 7.5, 7.7, 7.6, 7.4, 6.7, 7.2])
obs = mc_t_two_sample(I, II)
np.testing.assert_allclose(obs[:2], exp)
self.assertEqual(len(obs[2]), 999)
self.assertCorrectPValue(0.8, 0.9, mc_t_two_sample, [I, II],
p_val_idx=3)
def test_mc_t_two_sample_no_permuted_variance(self):
"""Test with chance of getting no variance with some perms."""
# Verified against R's t.test() and Deducer::perm.t.test().
x = np.array([1, 1, 2])
y = np.array([2, 2, 1])
exp = (-0.70710678118654791, 0.51851851851851838)
obs = mc_t_two_sample(x, y, permutations=10000)
np.testing.assert_allclose(obs[:2], exp)
self.assertEqual(len(obs[2]), 10000)
self.assertCorrectPValue(0.97, 1.0, mc_t_two_sample, [x, y],
{'permutations': 10000}, p_val_idx=3)
def test_mc_t_two_sample_no_permuted_variance(self):
"""Test with chance of getting no variance with some perms."""
# Verified against R's t.test() and Deducer::perm.t.test().
x = np.array([1, 1, 2])
y = np.array([2, 2, 1])
exp = (-0.70710678118654791, 0.51851851851851838)
obs = mc_t_two_sample(x, y, permutations=10000)
np.testing.assert_allclose(obs[:2], exp)
self.assertEqual(len(obs[2]), 10000)
self.assertCorrectPValue(0.97, 1.0, mc_t_two_sample, [x, y],
{'permutations': 10000}, p_val_idx=3)
def test_mc_t_two_sample_no_variance(self):
"""Test input with no variance. Should match Deducer::perm.t.test."""
x = np.array([1, 1, 1])
y = np.array([2, 2, 2])
exp = (float('-inf'), 0.0)
obs = mc_t_two_sample(x, y, permutations=10000)
self.assertEqual(obs[:2], exp)
self.assertEqual(len(obs[2]), 10000)
self.assertCorrectPValue(0.09, 0.11, mc_t_two_sample, [x, y],
{'permutations': 10000}, p_val_idx=3)
exp = (float('inf'), 0.0)
obs = mc_t_two_sample(y, x, permutations=10000)
self.assertEqual(obs[:2], exp)
self.assertEqual(len(obs[2]), 10000)
self.assertCorrectPValue(0.09, 0.11, mc_t_two_sample, [y, x],
{'permutations': 10000}, p_val_idx=3)
exp = (float('-inf'), 1.0)
obs = mc_t_two_sample(x, y, permutations=10000, tails='high')
self.assertEqual(obs[:2], exp)
self.assertEqual(len(obs[2]), 10000)
self.assertCorrectPValue(0.9999, 1.0, mc_t_two_sample, [x, y],
{'permutations': 10000, 'tails': 'high'},
p_val_idx=3)
exp = (float('-inf'), 0.0)
obs = mc_t_two_sample(x, y, permutations=10000, tails='low')
self.assertEqual(obs[:2], exp)
self.assertEqual(len(obs[2]), 10000)
self.assertCorrectPValue(0.04, 0.051, mc_t_two_sample, [x, y],
{'permutations': 10000, 'tails': 'low'},
p_val_idx=3)
def test_mc_t_two_sample_no_variance(self):
"""Test input with no variance. Should match Deducer::perm.t.test."""
x = np.array([1, 1, 1])
y = np.array([2, 2, 2])
exp = (float('-inf'), 0.0)
obs = mc_t_two_sample(x, y, permutations=10000)
self.assertEqual(obs[:2], exp)
self.assertEqual(len(obs[2]), 10000)
self.assertCorrectPValue(0.09, 0.11, mc_t_two_sample, [x, y],
{'permutations': 10000}, p_val_idx=3)
exp = (float('inf'), 0.0)
obs = mc_t_two_sample(y, x, permutations=10000)
self.assertEqual(obs[:2], exp)
self.assertEqual(len(obs[2]), 10000)
self.assertCorrectPValue(0.09, 0.11, mc_t_two_sample, [y, x],
{'permutations': 10000}, p_val_idx=3)
exp = (float('-inf'), 1.0)
obs = mc_t_two_sample(x, y, permutations=10000, tails='high')
self.assertEqual(obs[:2], exp)
self.assertEqual(len(obs[2]), 10000)
self.assertCorrectPValue(0.9999, 1.0, mc_t_two_sample, [x, y],
{'permutations': 10000, 'tails': 'high'},
p_val_idx=3)
exp = (float('-inf'), 0.0)
obs = mc_t_two_sample(x, y, permutations=10000, tails='low')
self.assertEqual(obs[:2], exp)
self.assertEqual(len(obs[2]), 10000)
self.assertCorrectPValue(0.04, 0.051, mc_t_two_sample, [x, y],
{'permutations': 10000, 'tails': 'low'},
p_val_idx=3)
def test_mc_t_two_sample_single_obs_sample(self):
"""Test works correctly with one sample having a single observation."""
sample = np.array([4.02, 3.88, 3.34, 3.87, 3.18])
x = np.array([3.02])
exp = (-1.5637254, 0.1929248)
obs = mc_t_two_sample(x, sample)
np.testing.assert_allclose(obs[:2], exp)
np.testing.assert_allclose(len(obs[2]), 999)
self.assertTrue(0.0 <= obs[3] <= 1.0)
exp = (1.5637254, 0.1929248)
obs = mc_t_two_sample(sample, x)
np.testing.assert_allclose(obs[:2], exp)
np.testing.assert_allclose(len(obs[2]), 999)
self.assertTrue(0.0 <= obs[3] <= 1.0)
# Test the case where we can have no variance in the permuted lists.
x = np.array([1, 1, 2])
y = np.array([1])
exp = (0.5, 0.666666666667)
obs = mc_t_two_sample(x, y)
np.testing.assert_allclose(obs[:2], exp)
np.testing.assert_allclose(len(obs[2]), 999)
self.assertTrue(0.0 <= obs[3] <= 1.0)
def test_mc_t_two_sample_single_obs_sample(self):
"""Test works correctly with one sample having a single observation."""
sample = np.array([4.02, 3.88, 3.34, 3.87, 3.18])
x = np.array([3.02])
exp = (-1.5637254, 0.1929248)
obs = mc_t_two_sample(x, sample)
np.testing.assert_allclose(obs[:2], exp)
np.testing.assert_allclose(len(obs[2]), 999)
self.assertTrue(0.0 <= obs[3] <= 1.0)
exp = (1.5637254, 0.1929248)
obs = mc_t_two_sample(sample, x)
np.testing.assert_allclose(obs[:2], exp)
np.testing.assert_allclose(len(obs[2]), 999)
self.assertTrue(0.0 <= obs[3] <= 1.0)
# Test the case where we can have no variance in the permuted lists.
x = np.array([1, 1, 2])
y = np.array([1])
exp = (0.5, 0.666666666667)
obs = mc_t_two_sample(x, y)
np.testing.assert_allclose(obs[:2], exp)
np.testing.assert_allclose(len(obs[2]), 999)
self.assertTrue(0.0 <= obs[3] <= 1.0)
def test_mc_t_two_sample_no_mc(self):
"""Test no MC stats if initial t-test is bad."""
x = np.array([1, 1, 1])
y = np.array([0, 0, 0])
self.assertEqual(mc_t_two_sample(x, x), (None, None, [], None))
def test_mc_t_two_sample_no_mc(self):
"""Test no MC stats if initial t-test is bad."""
x = np.array([1, 1, 1])
y = np.array([0, 0, 0])
self.assertEqual(mc_t_two_sample(x, x), (None, None, [], None))
