class TestJointProbabilityDistributionMethods(unittest.TestCase):
def setUp(self):
self.jpd = JPD(['x1', 'x2', 'x3'], [2, 3, 2], values=np.ones(12) / 12)
def test_jpd_marginal_distribution_list(self):
self.jpd.marginal_distribution(['x1', 'x2'])
np_test.assert_array_almost_equal(
self.jpd.values,
np.array([
0.16666667, 0.16666667, 0.16666667, 0.16666667, 0.16666667,
0.16666667
]))
np_test.assert_array_equal(self.jpd.cardinality, np.array([2, 3]))
dic = {
'x1': [('x1', 0), ('x1', 1)],
'x2': [('x2', 0), ('x2', 1), ('x2', 2)]
}
self.assertEqual(self.jpd.variables,
OrderedDict(sorted(dic.items(), key=lambda t: t[1])))
np_test.assert_almost_equal(np.sum(self.jpd.values), 1)
def test_marginal_distribution_str(self):
self.jpd.marginal_distribution('x1')
np_test.assert_array_almost_equal(self.jpd.values, np.array([0.5,
0.5]))
np_test.assert_array_equal(self.jpd.cardinality, np.array([2]))
dic = {'x1': [('x1', 0), ('x1', 1)]}
self.assertEqual(self.jpd.variables,
OrderedDict(sorted(dic.items(), key=lambda t: t[1])))
np_test.assert_almost_equal(np.sum(self.jpd.values), 1)
def test_conditional_distribution_list(self):
self.jpd.conditional_distribution([('x1', 1), ('x2', 0)])
np_test.assert_array_almost_equal(self.jpd.values, np.array([0.5,
0.5]))
np_test.assert_array_equal(self.jpd.cardinality, np.array([2]))
dic = {'x3': [('x3', 0), ('x3', 1)]}
self.assertEqual(self.jpd.variables,
OrderedDict(sorted(dic.items(), key=lambda t: t[1])))
np_test.assert_almost_equal(np.sum(self.jpd.values), 1)
def test_conditional_distribution_str(self):
self.jpd.conditional_distribution(('x1', 1))
np_test.assert_array_almost_equal(
self.jpd.values,
np.array([
0.16666667, 0.16666667, 0.16666667, 0.16666667, 0.16666667,
0.16666667
]))
np_test.assert_array_equal(self.jpd.cardinality, np.array([3, 2]))
dic = {
'x2': [('x2', 0), ('x2', 1), ('x2', 2)],
'x3': [('x3', 0), ('x3', 1)]
}
self.assertEqual(self.jpd.variables,
OrderedDict(sorted(dic.items(), key=lambda t: t[1])))
np_test.assert_almost_equal(np.sum(self.jpd.values), 1)
def tearDown(self):
del self.jpd
def test_is_imap(self):
val = [
0.01, 0.01, 0.08, 0.006, 0.006, 0.048, 0.004, 0.004, 0.032, 0.04,
0.04, 0.32, 0.024, 0.024, 0.192, 0.016, 0.016, 0.128
]
JPD = JointProbabilityDistribution(['diff', 'intel', 'grade'],
[2, 3, 3], val)
fac = Factor(['diff', 'intel', 'grade'], [2, 3, 3], val)
self.assertTrue(self.G1.is_imap(JPD))
self.assertRaises(TypeError, self.G1.is_imap, fac)
class TestJointProbabilityDistributionMethods(unittest.TestCase):
def setUp(self):
self.jpd = JPD(['x1', 'x2', 'x3'], [2, 3, 2], values=np.ones(12) / 12)
def test_jpd_marginal_distribution_list(self):
self.jpd.marginal_distribution(['x1', 'x2'])
np_test.assert_array_almost_equal(self.jpd.values, np.array([0.16666667, 0.16666667, 0.16666667,
0.16666667, 0.16666667, 0.16666667]))
np_test.assert_array_equal(self.jpd.cardinality, np.array([2, 3]))
dic = {'x1': ['x1_0', 'x1_1'], 'x2': ['x2_0', 'x2_1', 'x2_2']}
self.assertEqual(self.jpd.variables, OrderedDict(sorted(dic.items(), key=lambda t: t[1])))
np_test.assert_almost_equal(np.sum(self.jpd.values), 1)
def test_marginal_distribution_str(self):
self.jpd.marginal_distribution('x1')
np_test.assert_array_almost_equal(self.jpd.values, np.array([0.5, 0.5]))
np_test.assert_array_equal(self.jpd.cardinality, np.array([2]))
dic = {'x1': ['x1_0', 'x1_1']}
self.assertEqual(self.jpd.variables, OrderedDict(sorted(dic.items(), key=lambda t: t[1])))
np_test.assert_almost_equal(np.sum(self.jpd.values), 1)
def test_conditional_distribution_list(self):
self.jpd.conditional_distribution(['x1_1', 'x2_0'])
np_test.assert_array_almost_equal(self.jpd.values, np.array([0.5, 0.5]))
np_test.assert_array_equal(self.jpd.cardinality, np.array([2]))
dic = {'x3': ['x3_0', 'x3_1']}
self.assertEqual(self.jpd.variables, OrderedDict(sorted(dic.items(), key=lambda t: t[1])))
np_test.assert_almost_equal(np.sum(self.jpd.values), 1)
def test_conditional_distribution_str(self):
self.jpd.conditional_distribution('x1_1')
np_test.assert_array_almost_equal(self.jpd.values, np.array([0.16666667, 0.16666667,
0.16666667, 0.16666667,
0.16666667, 0.16666667]))
np_test.assert_array_equal(self.jpd.cardinality, np.array([3, 2]))
dic = {'x2': ['x2_0', 'x2_1', 'x2_2'], 'x3': ['x3_0', 'x3_1']}
self.assertEqual(self.jpd.variables, OrderedDict(sorted(dic.items(), key=lambda t: t[1])))
np_test.assert_almost_equal(np.sum(self.jpd.values), 1)
def tearDown(self):
del self.jpd
def setUp(self):
self.jpd = JPD(['x1', 'x2', 'x3'], [2, 3, 2], values=np.ones(12) / 12)
def setUp(self):
self.jpd = JPD(['x1', 'x2', 'x3'], [2, 3, 2], values=np.ones(12) / 12)
['A', 'B', 'C'])
independencies.get_assertations()
" (X _|_ Y) " " (A _|_ B|C) "
-- Joint Probability --
# --- Representing joint probability distributions using pgmpy
from pgmpy.factors import JointProbabilityDistribution as Joint
distribution = Joint(['coin1', 'coin2'], # names of the variables
[2, 2], # number of states of each variable
[0.25, 0.25, 0.25, 0.25]) # a list of probability values
distribution.check_independence('coin1', 'coin2')
" True "
-- Conditional Probability --
# --- Representing CPD in pgmpy
# Table CPD
def setUp(self):
self.jpd = JPD(['x1', 'x2', 'x3'], [2, 3, 2], values=np.ones(12) / 12)
self.jpd1 = JPD(['x1', 'x2', 'x3'], [2, 3, 2], values=np.ones(12) / 12)
self.jpd2 = JPD(['x1','x2','x3'],[2,2,3],
[0.126,0.168,0.126,0.009,0.045,0.126,0.252,0.0224,0.0056,0.06,0.036,0.024])
class TestJointProbabilityDistributionMethods(unittest.TestCase):
def setUp(self):
self.jpd = JPD(['x1', 'x2', 'x3'], [2, 3, 2], values=np.ones(12) / 12)
self.jpd1 = JPD(['x1', 'x2', 'x3'], [2, 3, 2], values=np.ones(12) / 12)
self.jpd2 = JPD(['x1','x2','x3'],[2,2,3],
[0.126,0.168,0.126,0.009,0.045,0.126,0.252,0.0224,0.0056,0.06,0.036,0.024])
def test_jpd_marginal_distribution_list(self):
self.jpd.marginal_distribution(['x1', 'x2'])
np_test.assert_array_almost_equal(self.jpd.values, np.array([[0.16666667, 0.16666667, 0.16666667],
[0.16666667, 0.16666667, 0.16666667]]))
np_test.assert_array_equal(self.jpd.cardinality, np.array([2, 3]))
dic = {'x1':2, 'x2':3}
self.assertEqual(self.jpd.get_cardinality(['x1','x2']), dic)
self.assertEqual(self.jpd.scope(), ['x1', 'x2'])
np_test.assert_almost_equal(np.sum(self.jpd.values), 1)
new_jpd = self.jpd1.marginal_distribution(['x1', 'x2'], inplace=False)
self.assertTrue(self.jpd1 != self.jpd)
self.assertTrue(new_jpd == self.jpd)
def test_marginal_distribution_str(self):
self.jpd.marginal_distribution('x1')
np_test.assert_array_almost_equal(self.jpd.values, np.array([0.5, 0.5]))
np_test.assert_array_equal(self.jpd.cardinality, np.array([2]))
self.assertEqual(self.jpd.scope(), ['x1'])
np_test.assert_almost_equal(np.sum(self.jpd.values), 1)
new_jpd = self.jpd1.marginal_distribution('x1', inplace=False)
self.assertTrue(self.jpd1 != self.jpd)
self.assertTrue(self.jpd == new_jpd)
def test_conditional_distribution_list(self):
self.jpd = self.jpd1.copy()
self.jpd.conditional_distribution([('x1', 1), ('x2', 0)])
np_test.assert_array_almost_equal(self.jpd.values, np.array([0.5, 0.5]))
np_test.assert_array_equal(self.jpd.cardinality, np.array([2]))
self.assertEqual(self.jpd.scope(), ['x3'])
np_test.assert_almost_equal(np.sum(self.jpd.values), 1)
new_jpd = self.jpd1.conditional_distribution([('x1', 1), ('x2', 0)], inplace=False)
self.assertTrue(self.jpd1 != self.jpd)
self.assertTrue(self.jpd == new_jpd)
def test_check_independence(self):
self.assertTrue(self.jpd2.check_independence(['x1'], ['x2']))
self.assertRaises(TypeError, self.jpd2.check_independence, 'x1', ['x2'])
self.assertRaises(TypeError, self.jpd2.check_independence, ['x1'], 'x2')
self.assertRaises(TypeError, self.jpd2.check_independence,['x1'], ['x2'], 'x3')
self.assertFalse(self.jpd2.check_independence(['x1'], ['x2'], ('x3',), condition_random_variable=True))
self.assertFalse(self.jpd2.check_independence(['x1'], ['x2'], [('x3', 0)]))
self.assertTrue(self.jpd1.check_independence(['x1'], ['x2'], ('x3',), condition_random_variable=True))
self.assertTrue(self.jpd1.check_independence(['x1'], ['x2'], [('x3', 1)]))
def test_get_independencies(self):
independencies = Independencies(['x1', 'x2'], ['x2', 'x3'], ['x3', 'x1'])
independencies1 = Independencies(['x1', 'x2'])
self.assertEqual(self.jpd1.get_independencies(), independencies)
self.assertEqual(self.jpd2.get_independencies(), independencies1)
self.assertEqual(self.jpd1.get_independencies([('x3', 0)]), independencies1)
self.assertEqual(self.jpd2.get_independencies([('x3', 0)]), Independencies())
def test_minimal_imap(self):
bm = self.jpd1.minimal_imap(order=['x1','x2','x3'])
self.assertEqual(sorted(bm.edges()), sorted([('x1', 'x3'), ('x2', 'x3')]))
bm = self.jpd1.minimal_imap(order=['x2','x3','x1'])
self.assertEqual(sorted(bm.edges()), sorted([('x2', 'x1'), ('x3', 'x1')]))
bm = self.jpd2.minimal_imap(order=['x1','x2','x3'])
self.assertEqual(bm.edges(), [])
bm = self.jpd2.minimal_imap(order=['x1','x2'])
self.assertEqual(bm.edges(), [])
def test_repr(self):
self.assertEqual(repr(self.jpd1),'<Joint Distribution representing P(x1:2, x2:3, x3:2) at {address}>'
.format(address=hex(id(self.jpd1))))
def tearDown(self):
del self.jpd
del self.jpd1
del self.jpd2
def test_jpd_init(self):
jpd = JPD(['x1', 'x2', 'x3'], [2, 3, 2], np.ones(12) / 12)
np_test.assert_array_equal(jpd.cardinality, np.array([2, 3, 2]))
np_test.assert_array_equal(jpd.values, np.ones(12).reshape(2, 3, 2) / 12)
self.assertEqual(jpd.get_cardinality(['x1', 'x2', 'x3']), {'x1':2, 'x2':3, 'x3':2})