def __init__(self, variables, cardinality, values):
"""
Initialize a Joint Probability Distribution class.
Defined above, we have the following mapping from variable
assignments to the index of the row vector in the value field:
+-----+-----+-----+-------------------------+
| x1 | x2 | x3 | P(x1, x2, x2) |
+-----+-----+-----+-------------------------+
| x1_0| x2_0| x3_0| P(x1_0, x2_0, x3_0) |
+-----+-----+-----+-------------------------+
| x1_1| x2_0| x3_0| P(x1_1, x2_0, x3_0) |
+-----+-----+-----+-------------------------+
| x1_0| x2_1| x3_0| P(x1_0, x2_1, x3_0) |
+-----+-----+-----+-------------------------+
| x1_1| x2_1| x3_0| P(x1_1, x2_1, x3_0) |
+-----+-----+-----+-------------------------+
| x1_0| x2_0| x3_1| P(x1_0, x2_0, x3_1) |
+-----+-----+-----+-------------------------+
| x1_1| x2_0| x3_1| P(x1_1, x2_0, x3_1) |
+-----+-----+-----+-------------------------+
| x1_0| x2_1| x3_1| P(x1_0, x2_1, x3_1) |
+-----+-----+-----+-------------------------+
| x1_1| x2_1| x3_1| P(x1_1, x2_1, x3_1) |
+-----+-----+-----+-------------------------+
Parameters
----------
variables: list
List of scope of Joint Probability Distribution.
cardinality: list, array_like
List of cardinality of each variable
value: list, array_like
List or array of values of factor.
A Joint Probability Distribution's values are stored in a row
vector in the value using an ordering such that the left-most
variables as defined in the variable field cycle through their
values the fastest.
Examples
--------
>>> from pgmpy.factors import JointProbabilityDistribution
>>> prob = JointProbabilityDistribution(['x1', 'x2', 'x3'], [2, 2, 2], np.ones(8)/8)
>>> print(prob)
print(prob)
x1 x2 x3 P(x1, x2, x3)
x1_0 x2_0 x3_0 0.125
x1_0 x2_0 x3_1 0.125
x1_0 x2_1 x3_0 0.125
x1_0 x2_1 x3_1 0.125
x1_1 x2_0 x3_0 0.125
x1_1 x2_0 x3_1 0.125
x1_1 x2_1 x3_0 0.125
x1_1 x2_1 x3_1 0.125
"""
if np.isclose(np.sum(values), 1):
Factor.__init__(self, variables, cardinality, values)
else:
raise ValueError("The probability values doesn't sum to 1.")
def test_factor_product2(self):
from pgmpy import factors
phi = factors.Factor(['x1', 'x2'], [2, 2], range(4))
phi1 = factors.Factor(['x3', 'x4'], [2, 2], range(4))
prod = phi.product(phi1)
np_test.assert_array_equal(prod.values,
np.array([0, 0, 0, 0, 0, 1,
2, 3, 0, 2, 4, 6,
0, 3, 6, 9]))
self.assertEqual(prod.variables, OrderedDict([
('x1', [State('x1', 0), State('x1', 1)]),
('x2', [State('x2', 0), State('x2', 1)]),
('x3', [State('x3', 0), State('x3', 1)]),
('x4', [State('x4', 0), State('x4', 1)])]
))
phi = Factor(['x1', 'x2'], [3, 2], range(6))
phi1 = Factor(['x2', 'x3'], [2, 2], range(4))
prod = phi.product(phi1)
np_test.assert_array_equal(prod.values,
np.array([0, 0, 2, 3, 0, 2, 6, 9, 0, 4, 10, 15]))
self.assertEqual(prod.variables, OrderedDict(
[('x1', [State('x1', 0), State('x1', 1), State('x1', 2)]),
('x2', [State('x2', 0), State('x2', 1)]),
('x3', [State('x3', 0), State('x3', 1)])]))
def __init__(self, variable, variable_card, values,
evidence=None, evidence_card=None):
self.variable = variable
variables = [variable]
if not isinstance(variable_card, int):
raise TypeError("Event cardinality must be an integer")
self.variable_card = variable_card
cardinality = [variable_card]
if evidence_card:
if not isinstance(evidence_card, (list, set, tuple)):
evidence_card = [evidence_card]
cardinality.extend(evidence_card)
if evidence:
if not isinstance(evidence, (list, set, tuple)):
evidence = [evidence]
variables.extend(evidence)
if not len(evidence_card) == len(evidence):
raise exceptions.CardinalityError("Cardinality of all "
"evidences not specified")
values = np.array(values)
if values.ndim != 2:
raise TypeError("Values must be a 2d list/array")
Factor.__init__(self, variables, cardinality, values.flatten('C'))
def setUp(self):
self.phi = Factor(['x1', 'x2', 'x3'], [2, 2, 2], np.random.uniform(5, 10, size=8))
self.phi1 = Factor(['x1', 'x2', 'x3'], [2, 3, 2], range(12))
self.phi2 = Factor([('x1', 0), ('x2', 0), ('x3', 0)], [2, 3, 2], range(12))
# This larger factor (phi3) caused a bug in reduce
card3 = [3, 3, 3, 2, 2, 2, 2, 2, 2]
self.phi3 = Factor(['A', 'B', 'C', 'D', 'E', 'F', 'G', 'H', 'I'],
card3, np.arange(np.prod(card3), dtype=np.float))
def test_factorset_divide(self):
phi1 = Factor(['x1', 'x2', 'x3'], [2, 3, 2], range(1, 13))
phi2 = Factor(['x3', 'x4', 'x1'], [2, 2, 2], range(1, 9))
factor_set1 = FactorSet(phi1, phi2)
phi3 = Factor(['x5', 'x6', 'x7'], [2, 2, 2], range(1, 9))
phi4 = Factor(['x5', 'x7', 'x8'], [2, 2, 2], range(1, 9))
factor_set2 = FactorSet(phi3, phi4)
factor_set3 = factor_set2.divide(factor_set1, inplace=False)
self.assertEqual({phi3, phi4, phi1.identity_factor() / phi1, phi2.identity_factor() / phi2},
factor_set3.factors)
class TimeSuite:
def setup(self):
self.phi = Factor(['x1', 'x2', 'x3'], [2, 2, 2], np.ones(8))
self.phi_large = Factor(range(10), [2] * 10, [1] * (2**10))
def time_factor_define(self):
phi = Factor(['x1', 'x2', 'x3'], [2, 2, 2], np.ones(8))
def time_factor_reduce(self):
self.phi.reduce([('x1', 0), ('x2', 1)], inplace=False)
def time_factor_reduce_large(self):
self.phi_large.reduce([(3, 0), (6, 1)], inplace=False)
def time_factor_marginalize(self):
self.phi.marginalize(['x1'], inplace=False)
def time_factor_marginalize_large(self):
self.phi_large.marginalize([4, 5, 8], inplace=False)
def time_factor_multiply(self):
phi = self.phi * self.phi
def time_factor_multiply_large(self):
phi = self.phi_large * self.phi_large
def test_max_calibrate_clique_belief(self):
belief_propagation = BeliefPropagation(self.junction_tree)
belief_propagation.max_calibrate()
clique_belief = belief_propagation.get_clique_beliefs()
phi1 = Factor(['A', 'B'], [2, 3], range(6))
phi2 = Factor(['B', 'C'], [3, 2], range(6))
phi3 = Factor(['C', 'D'], [2, 2], range(4))
b_A_B = phi1 * (phi3.maximize(['D'], inplace=False) * phi2).maximize(['C'], inplace=False)
b_B_C = phi2 * (phi1.maximize(['A'], inplace=False) * phi3.maximize(['D'], inplace=False))
b_C_D = phi3 * (phi1.maximize(['A'], inplace=False) * phi2).maximize(['B'], inplace=False)
np_test.assert_array_almost_equal(clique_belief[('A', 'B')].values, b_A_B.values)
np_test.assert_array_almost_equal(clique_belief[('B', 'C')].values, b_B_C.values)
np_test.assert_array_almost_equal(clique_belief[('C', 'D')].values, b_C_D.values)
def test_factor_product2(self):
from pgmpy import factors
phi = factors.Factor(['x1', 'x2'], [2, 2], range(4))
phi1 = factors.Factor(['x3', 'x4'], [2, 2], range(4))
prod = phi.product(phi1, inplace=False)
expected_factor = Factor(['x1', 'x2', 'x3', 'x4'], [2, 2, 2, 2],
[0, 0, 0, 0, 0, 1, 2, 3, 0, 2, 4, 6, 0, 3, 6, 9])
self.assertEqual(prod, expected_factor)
self.assertEqual(sorted(prod.variables), ['x1', 'x2', 'x3', 'x4'])
phi = Factor(['x1', 'x2'], [3, 2], range(6))
phi1 = Factor(['x2', 'x3'], [2, 2], range(4))
prod = phi.product(phi1, inplace=False)
expected_factor = Factor(['x1', 'x2', 'x3'], [3, 2, 2],
[0, 0, 2, 3, 0, 2, 6, 9, 0, 4, 10, 15])
self.assertEqual(prod, expected_factor)
self.assertEqual(sorted(prod.variables), ['x1', 'x2', 'x3'])
def test_max_calibrate_sepset_belief(self):
belief_propagation = BeliefPropagation(self.junction_tree)
belief_propagation.max_calibrate()
sepset_belief = belief_propagation.get_sepset_beliefs()
phi1 = Factor(['A', 'B'], [2, 3], range(6))
phi2 = Factor(['B', 'C'], [3, 2], range(6))
phi3 = Factor(['C', 'D'], [2, 2], range(4))
b_B = (phi1 * (phi3.maximize(['D'], inplace=False) *
phi2).maximize(['C'], inplace=False)).maximize(['A'], inplace=False)
b_C = (phi2 * (phi1.maximize(['A'], inplace=False) *
phi3.maximize(['D'], inplace=False))).maximize(['B'], inplace=False)
np_test.assert_array_almost_equal(sepset_belief[frozenset((('A', 'B'), ('B', 'C')))].values, b_B.values)
np_test.assert_array_almost_equal(sepset_belief[frozenset((('B', 'C'), ('C', 'D')))].values, b_C.values)
def test_factorset_marginalize_not_inplace(self):
phi1 = Factor(['x1', 'x2', 'x3'], [2, 3, 2], range(12))
phi2 = Factor(['x3', 'x4', 'x1'], [2, 2, 2], range(8))
phi3 = Factor(['x5', 'x6', 'x7'], [2, 2, 2], range(8))
phi4 = Factor(['x5', 'x7', 'x8'], [2, 2, 2], range(8))
factor_set = FactorSet(phi1, phi2, phi3, phi4)
new_factor_set = factor_set.marginalize(['x1', 'x5'], inplace=False)
phi1_equivalent_in_factor_set = list(filter(lambda x: set(x.scope()) == set(['x2', 'x3']),
new_factor_set.factors))[0]
self.assertEqual(phi1.marginalize('x1', inplace=False), phi1_equivalent_in_factor_set)
phi2_equivalent_in_factor_set = list(filter(lambda x: set(x.scope()) == set(['x4', 'x3']),
new_factor_set.factors))[0]
self.assertEqual(phi2.marginalize('x1', inplace=False), phi2_equivalent_in_factor_set)
phi3_equivalent_in_factor_set = list(filter(lambda x: set(x.scope()) == set(['x6', 'x7']),
new_factor_set.factors))[0]
self.assertEqual(phi3.marginalize('x5', inplace=False), phi3_equivalent_in_factor_set)
phi4_equivalent_in_factor_set = list(filter(lambda x: set(x.scope()) == set(['x8', 'x7']),
new_factor_set.factors))[0]
self.assertEqual(phi4.marginalize('x5', inplace=False), phi4_equivalent_in_factor_set)
class TestFactorMethods(unittest.TestCase):
def setUp(self):
self.phi = Factor(['x1', 'x2', 'x3'], [2, 2, 2], np.random.uniform(5, 10, size=8))
self.phi1 = Factor(['x1', 'x2', 'x3'], [2, 3, 2], range(12))
self.phi2 = Factor([('x1', 0), ('x2', 0), ('x3', 0)], [2, 3, 2], range(12))
# This larger factor (phi3) caused a bug in reduce
card3 = [3, 3, 3, 2, 2, 2, 2, 2, 2]
self.phi3 = Factor(['A', 'B', 'C', 'D', 'E', 'F', 'G', 'H', 'I'],
card3, np.arange(np.prod(card3), dtype=np.float))
def test_scope(self):
self.assertListEqual(self.phi.scope(), ['x1', 'x2', 'x3'])
self.assertListEqual(self.phi1.scope(), ['x1', 'x2', 'x3'])
def test_assignment(self):
self.assertListEqual(self.phi.assignment([0]), [[('x1', 0), ('x2', 0), ('x3', 0)]])
self.assertListEqual(self.phi.assignment([4, 5, 6]), [[('x1', 1), ('x2', 0), ('x3', 0)],
[('x1', 1), ('x2', 0), ('x3', 1)],
[('x1', 1), ('x2', 1), ('x3', 0)]])
self.assertListEqual(self.phi1.assignment(np.array([4, 5, 6])),
[[('x1', 0), ('x2', 2), ('x3', 0)],
[('x1', 0), ('x2', 2), ('x3', 1)],
[('x1', 1), ('x2', 0), ('x3', 0)]])
def test_assignment_indexerror(self):
self.assertRaises(IndexError, self.phi.assignment, [10])
self.assertRaises(IndexError, self.phi.assignment, [1, 3, 10, 5])
self.assertRaises(IndexError, self.phi.assignment, np.array([1, 3, 10, 5]))
def test_get_cardinality(self):
self.assertEqual(self.phi.get_cardinality(['x1']), {'x1': 2})
self.assertEqual(self.phi.get_cardinality(['x2']), {'x2': 2})
self.assertEqual(self.phi.get_cardinality(['x3']), {'x3': 2})
self.assertEqual(self.phi.get_cardinality(['x1', 'x2']), {'x1': 2, 'x2': 2})
self.assertEqual(self.phi.get_cardinality(['x1', 'x3']), {'x1': 2, 'x3': 2})
self.assertEqual(self.phi.get_cardinality(['x1', 'x2', 'x3']), {'x1': 2, 'x2': 2, 'x3': 2})
def test_get_cardinality_scopeerror(self):
self.assertRaises(ValueError, self.phi.get_cardinality, ['x4'])
def test_get_cardinality_typeerror(self):
self.assertRaises(TypeError, self.phi.get_cardinality, 'x1')
def test_marginalize(self):
self.phi1.marginalize(['x1'])
np_test.assert_array_equal(self.phi1.values, np.array([[6, 8],
[10, 12],
[14, 16]]))
self.phi1.marginalize(['x2'])
np_test.assert_array_equal(self.phi1.values, np.array([30, 36]))
self.phi1.marginalize(['x3'])
np_test.assert_array_equal(self.phi1.values, np.array(66))
def test_marginalize_scopeerror(self):
self.assertRaises(ValueError, self.phi.marginalize, ['x4'])
self.assertRaises(ValueError, self.phi.marginalize, ['x4'])
self.phi.marginalize(['x1'])
self.assertRaises(ValueError, self.phi.marginalize, ['x1'])
def test_marginalize_typeerror(self):
self.assertRaises(TypeError, self.phi.marginalize, 'x1')
def test_marginalize_shape(self):
values = ['A', 'D', 'F', 'H']
phi3_max = self.phi3.marginalize(values, inplace=False)
# Previously a sorting error caused these to be different
np_test.assert_array_equal(phi3_max.values.shape, phi3_max.cardinality)
def test_normalize(self):
self.phi1.normalize()
np_test.assert_almost_equal(self.phi1.values,
np.array([[[0, 0.01515152],
[0.03030303, 0.04545455],
[0.06060606, 0.07575758]],
[[0.09090909, 0.10606061],
[0.12121212, 0.13636364],
[0.15151515, 0.16666667]]]))
def test_reduce(self):
self.phi1.reduce([('x1', 0), ('x2', 0)])
np_test.assert_array_equal(self.phi1.values, np.array([0, 1]))
def test_reduce1(self):
self.phi1.reduce([('x2', 0), ('x1', 0)])
np_test.assert_array_equal(self.phi1.values, np.array([0, 1]))
def test_reduce_shape(self):
values = [('A', 0), ('D', 0), ('F', 0), ('H', 1)]
phi3_reduced = self.phi3.reduce(values, inplace=False)
# Previously a sorting error caused these to be different
np_test.assert_array_equal(phi3_reduced.values.shape, phi3_reduced.cardinality)
@unittest.skip
def test_complete_reduce(self):
self.phi1.reduce([('x1', 0), ('x2', 0), ('x3', 1)])
np_test.assert_array_equal(self.phi1.values, np.array([1]))
np_test.assert_array_equal(self.phi1.cardinality, np.array([]))
np_test.assert_array_equal(self.phi1.variables, OrderedDict())
def test_reduce_typeerror(self):
self.assertRaises(TypeError, self.phi1.reduce, 'x10')
self.assertRaises(TypeError, self.phi1.reduce, ['x10'])
self.assertRaises(TypeError, self.phi1.reduce, [('x1', 'x2')])
self.assertRaises(TypeError, self.phi1.reduce, [(0, 'x1')])
self.assertRaises(TypeError, self.phi1.reduce, [(0.1, 'x1')])
self.assertRaises(TypeError, self.phi1.reduce, [(0.1, 0.1)])
self.assertRaises(TypeError, self.phi1.reduce, [('x1', 0.1)])
def test_reduce_scopeerror(self):
self.assertRaises(ValueError, self.phi1.reduce, [('x4', 1)])
def test_reduce_sizeerror(self):
self.assertRaises(IndexError, self.phi1.reduce, [('x3', 5)])
def test_identity_factor(self):
identity_factor = self.phi.identity_factor()
self.assertEqual(list(identity_factor.variables), ['x1', 'x2', 'x3'])
np_test.assert_array_equal(identity_factor.cardinality, [2, 2, 2])
np_test.assert_array_equal(identity_factor.values, np.ones(8).reshape(2, 2, 2))
def test_factor_product(self):
phi = Factor(['x1', 'x2'], [2, 2], range(4))
phi1 = Factor(['x3', 'x4'], [2, 2], range(4))
prod = factor_product(phi, phi1)
expected_factor = Factor(['x1', 'x2', 'x3', 'x4'], [2, 2, 2, 2], [0, 0, 0, 0, 0, 1,
2, 3, 0, 2, 4, 6,
0, 3, 6, 9])
self.assertEqual(prod, expected_factor)
self.assertEqual(sorted(prod.variables), ['x1', 'x2', 'x3', 'x4'])
phi = Factor(['x1', 'x2'], [3, 2], range(6))
phi1 = Factor(['x2', 'x3'], [2, 2], range(4))
prod = factor_product(phi, phi1)
expected_factor = Factor(['x1', 'x2', 'x3'], [3, 2, 2], [0, 0, 2, 3, 0, 2,
6, 9, 0, 4, 10, 15])
np_test.assert_almost_equal(prod.values,
np.array([0, 0, 2, 3, 0, 2,
6, 9, 0, 4, 10, 15]).reshape(3, 2, 2))
self.assertEqual(sorted(prod.variables), ['x1', 'x2', 'x3'])
def test_factor_product2(self):
from pgmpy import factors
phi = factors.Factor(['x1', 'x2'], [2, 2], range(4))
phi1 = factors.Factor(['x3', 'x4'], [2, 2], range(4))
prod = phi.product(phi1, inplace=False)
expected_factor = Factor(['x1', 'x2', 'x3', 'x4'], [2, 2, 2, 2],
[0, 0, 0, 0, 0, 1, 2, 3, 0, 2, 4, 6, 0, 3, 6, 9])
self.assertEqual(prod, expected_factor)
self.assertEqual(sorted(prod.variables), ['x1', 'x2', 'x3', 'x4'])
phi = Factor(['x1', 'x2'], [3, 2], range(6))
phi1 = Factor(['x2', 'x3'], [2, 2], range(4))
prod = phi.product(phi1, inplace=False)
expected_factor = Factor(['x1', 'x2', 'x3'], [3, 2, 2],
[0, 0, 2, 3, 0, 2, 6, 9, 0, 4, 10, 15])
self.assertEqual(prod, expected_factor)
self.assertEqual(sorted(prod.variables), ['x1', 'x2', 'x3'])
def test_factor_product_non_factor_arg(self):
self.assertRaises(TypeError, factor_product, 1, 2)
def test_factor_mul(self):
phi = Factor(['x1', 'x2'], [2, 2], range(4))
phi1 = Factor(['x3', 'x4'], [2, 2], range(4))
prod = phi * phi1
sorted_vars = ['x1', 'x2', 'x3', 'x4']
for axis in range(prod.values.ndim):
exchange_index = prod.variables.index(sorted_vars[axis])
prod.variables[axis], prod.variables[exchange_index] = prod.variables[exchange_index], prod.variables[axis]
prod.values = prod.values.swapaxes(axis, exchange_index)
np_test.assert_almost_equal(prod.values.ravel(),
np.array([0, 0, 0, 0, 0, 1,
2, 3, 0, 2, 4, 6,
0, 3, 6, 9]))
self.assertEqual(prod.variables, ['x1', 'x2', 'x3', 'x4'])
def test_factor_divide(self):
phi1 = Factor(['x1', 'x2'], [2, 2], [1, 2, 2, 4])
phi2 = Factor(['x1'], [2], [1, 2])
div = phi1.divide(phi2, inplace=False)
phi3 = Factor(['x1', 'x2'], [2, 2], [1, 2, 1, 2])
self.assertEqual(phi3, div)
def test_factor_divide_truediv(self):
phi1 = Factor(['x1', 'x2'], [2, 2], [1, 2, 2, 4])
phi2 = Factor(['x1'], [2], [1, 2])
div = phi1 / phi2
phi3 = Factor(['x1', 'x2'], [2, 2], [1, 2, 1, 2])
self.assertEqual(phi3, div)
def test_factor_divide_invalid(self):
phi1 = Factor(['x1', 'x2'], [2, 2], [1, 2, 3, 4])
phi2 = Factor(['x1'], [2], [0, 2])
div = phi1.divide(phi2, inplace=False)
np_test.assert_array_equal(div.values.ravel(), np.array([np.inf, np.inf, 1.5, 2]))
def test_factor_divide_no_common_scope(self):
phi1 = Factor(['x1', 'x2'], [2, 2], [1, 2, 3, 4])
phi2 = Factor(['x3'], [2], [0, 2])
self.assertRaises(ValueError, factor_divide, phi1, phi2)
def test_factor_divide_non_factor_arg(self):
self.assertRaises(TypeError, factor_divide, 1, 1)
def test_eq(self):
self.assertFalse(self.phi == self.phi1)
self.assertTrue(self.phi == self.phi)
self.assertTrue(self.phi1 == self.phi1)
def test_eq1(self):
phi1 = Factor(['x1', 'x2', 'x3'], [2, 4, 3], range(24))
phi2 = Factor(['x2', 'x1', 'x3'], [4, 2, 3], [0, 1, 2, 12, 13, 14, 3,
4, 5, 15, 16, 17, 6, 7,
8, 18, 19, 20, 9, 10, 11,
21, 22, 23])
self.assertTrue(phi1, phi2)
def test_maximize1(self):
self.phi1.maximize(['x1'])
self.assertEqual(self.phi1, Factor(['x2', 'x3'], [3, 2], [6, 7, 8, 9, 10, 11]))
self.phi1.maximize(['x2'])
self.assertEqual(self.phi1, Factor(['x3'], [2], [10, 11]))
def test_maximize2(self):
self.phi1.maximize(['x1', 'x2'])
self.assertEqual(self.phi1, Factor(['x3'], [2], [10, 11]))
def test_maximize3(self):
self.phi2 = Factor(['x1', 'x2', 'x3'], [3, 2, 2], [0.25, 0.35, 0.08, 0.16, 0.05, 0.07,
0.00, 0.00, 0.15, 0.21, 0.08, 0.18])
self.phi2.maximize(['x2'])
self.assertEqual(self.phi2, Factor(['x1', 'x3'], [3, 2], [0.25, 0.35, 0.05,
0.07, 0.15, 0.21]))
def test_maximize_shape(self):
values = ['A', 'D', 'F', 'H']
phi3_max = self.phi3.maximize(values, inplace=False)
# Previously a sorting error caused these to be different
np_test.assert_array_equal(phi3_max.values.shape, phi3_max.cardinality)
def test_maximize_scopeerror(self):
self.assertRaises(ValueError, self.phi.maximize, ['x10'])
def test_maximize_typeerror(self):
self.assertRaises(TypeError, self.phi.maximize, 'x1')
def tearDown(self):
del self.phi
del self.phi1
def test_maximize3(self):
self.phi2 = Factor(['x1', 'x2', 'x3'], [3, 2, 2], [0.25, 0.35, 0.08, 0.16, 0.05, 0.07,
0.00, 0.00, 0.15, 0.21, 0.08, 0.18])
self.phi2.maximize(['x2'])
self.assertEqual(self.phi2, Factor(['x1', 'x3'], [3, 2], [0.25, 0.35, 0.05,
0.07, 0.15, 0.21]))
def test_factor_divide_invalid(self):
phi1 = Factor(['x1', 'x2'], [2, 2], [1, 2, 3, 4])
phi2 = Factor(['x1'], [2], [0, 2])
div = phi1.divide(phi2, inplace=False)
np_test.assert_array_equal(div.values.ravel(), np.array([np.inf, np.inf, 1.5, 2]))
def test_factor_divide(self):
phi1 = Factor(['x1', 'x2'], [2, 2], [1, 2, 2, 4])
phi2 = Factor(['x1'], [2], [1, 2])
div = phi1.divide(phi2, inplace=False)
phi3 = Factor(['x1', 'x2'], [2, 2], [1, 2, 1, 2])
self.assertEqual(phi3, div)
from pgmpy.factors import Factor
factor_a_b = Factor(variables=['A', 'B'],
cardinality=[2, 2],
value=[90, 100, 1, 10])
factor_b_c = Factor(variables=['B', 'C'],
cardinality=[2, 2],
value=[10, 80, 70, 30])
factor_c_d = Factor(variables=['C', 'D'],
cardinality=[2, 2],
value=[10, 1, 100, 90])
factor_d_a = Factor(variables=['D', 'A'],
cardinality=[2, 2],
value=[80, 60, 20, 10])
class TestFactorSet(unittest.TestCase):
def setUp(self):
self.phi1 = Factor(['x1', 'x2', 'x3'], [2, 3, 2], range(12))
self.phi2 = Factor(['x3', 'x4', 'x1'], [2, 2, 2], range(8))
self.phi3 = Factor(['x5', 'x6', 'x7'], [2, 2, 2], range(8))
self.phi4 = Factor(['x5', 'x7', 'x8'], [2, 2, 2], range(8))
def test_class_init(self):
phi1 = Factor(['x1', 'x2', 'x3'], [2, 3, 2], range(12))
phi2 = Factor(['x3', 'x4', 'x1'], [2, 2, 2], range(8))
factor_set1 = FactorSet(phi1, phi2)
self.assertEqual({phi1, phi2}, factor_set1.get_factors())
def test_factorset_add_remove_factors(self):
self.factor_set1 = FactorSet()
self.factor_set1.add_factors(self.phi1, self.phi2)
self.assertEqual({self.phi1, self.phi2}, self.factor_set1.get_factors())
self.factor_set1.remove_factors(self.phi2)
self.assertEqual({self.phi1}, self.factor_set1.get_factors())
def test_factorset_product(self):
factor_set1 = FactorSet(self.phi1, self.phi2)
factor_set2 = FactorSet(self.phi3, self.phi4)
factor_set3 = factor_set2.product(factor_set1, inplace=False)
self.assertEqual({self.phi1, self.phi2, self.phi3, self.phi4}, factor_set3.factors)
def test_factorset_divide(self):
phi1 = Factor(['x1', 'x2', 'x3'], [2, 3, 2], range(1, 13))
phi2 = Factor(['x3', 'x4', 'x1'], [2, 2, 2], range(1, 9))
factor_set1 = FactorSet(phi1, phi2)
phi3 = Factor(['x5', 'x6', 'x7'], [2, 2, 2], range(1, 9))
phi4 = Factor(['x5', 'x7', 'x8'], [2, 2, 2], range(1, 9))
factor_set2 = FactorSet(phi3, phi4)
factor_set3 = factor_set2.divide(factor_set1, inplace=False)
self.assertEqual({phi3, phi4, phi1.identity_factor() / phi1, phi2.identity_factor() / phi2},
factor_set3.factors)
def test_factorset_marginalize_inplace(self):
factor_set = FactorSet(self.phi1, self.phi2, self.phi3, self.phi4)
factor_set.marginalize(['x1', 'x5'], inplace=True)
phi1_equivalent_in_factor_set = list(filter(lambda x: set(x.scope()) == {'x2', 'x3'},
factor_set.factors))[0]
self.assertEqual(self.phi1.marginalize(['x1'], inplace=False), phi1_equivalent_in_factor_set)
phi2_equivalent_in_factor_set = list(filter(lambda x: set(x.scope()) == {'x4', 'x3'},
factor_set.factors))[0]
self.assertEqual(self.phi2.marginalize(['x1'], inplace=False), phi2_equivalent_in_factor_set)
phi3_equivalent_in_factor_set = list(filter(lambda x: set(x.scope()) == {'x6', 'x7'},
factor_set.factors))[0]
self.assertEqual(self.phi3.marginalize(['x5'], inplace=False), phi3_equivalent_in_factor_set)
phi4_equivalent_in_factor_set = list(filter(lambda x: set(x.scope()) == {'x8', 'x7'},
factor_set.factors))[0]
self.assertEqual(self.phi4.marginalize(['x5'], inplace=False), phi4_equivalent_in_factor_set)
def test_factorset_marginalize_not_inplace(self):
factor_set = FactorSet(self.phi1, self.phi2, self.phi3, self.phi4)
new_factor_set = factor_set.marginalize(['x1', 'x5'], inplace=False)
phi1_equivalent_in_factor_set = list(filter(lambda x: set(x.scope()) == {'x2', 'x3'},
new_factor_set.factors))[0]
self.assertEqual(self.phi1.marginalize(['x1'], inplace=False), phi1_equivalent_in_factor_set)
phi2_equivalent_in_factor_set = list(filter(lambda x: set(x.scope()) == {'x4', 'x3'},
new_factor_set.factors))[0]
self.assertEqual(self.phi2.marginalize(['x1'], inplace=False), phi2_equivalent_in_factor_set)
phi3_equivalent_in_factor_set = list(filter(lambda x: set(x.scope()) == {'x6', 'x7'},
new_factor_set.factors))[0]
self.assertEqual(self.phi3.marginalize(['x5'], inplace=False), phi3_equivalent_in_factor_set)
phi4_equivalent_in_factor_set = list(filter(lambda x: set(x.scope()) == {'x8', 'x7'},
new_factor_set.factors))[0]
self.assertEqual(self.phi4.marginalize(['x5'], inplace=False), phi4_equivalent_in_factor_set)
class TestFactorMethods(unittest.TestCase):
def setUp(self):
self.phi = Factor(['x1', 'x2', 'x3'], [2, 2, 2], np.random.uniform(5, 10, size=8))
self.phi1 = Factor(['x1', 'x2', 'x3'], [2, 3, 2], range(12))
# def test_pos_dist(self):
# self.assertTrue(self.phi.is_pos_dist())
# self.assertFalse(self.phi1.is_pos_dist())
def test_scope(self):
self.assertListEqual(self.phi.scope(), ['x1', 'x2', 'x3'])
self.assertListEqual(self.phi1.scope(), ['x1', 'x2', 'x3'])
def test_assignment(self):
self.assertListEqual(self.phi.assignment([0]), [['x1_0', 'x2_0', 'x3_0']])
self.assertListEqual(self.phi.assignment([4, 5, 6]), [['x1_1', 'x2_0', 'x3_0'],
['x1_1', 'x2_0', 'x3_1'],
['x1_1', 'x2_1', 'x3_0']])
self.assertListEqual(self.phi1.assignment(np.array([4, 5, 6])), [['x1_0', 'x2_2', 'x3_0'],
['x1_0', 'x2_2', 'x3_1'],
['x1_1', 'x2_0', 'x3_0']])
def test_assignment_indexerror(self):
self.assertRaises(IndexError, self.phi.assignment, [10])
self.assertRaises(IndexError, self.phi.assignment, [1, 3, 10, 5])
self.assertRaises(IndexError, self.phi.assignment, np.array([1, 3, 10, 5]))
# def test_get_variables(self):
# self.assertListEqual(sorted(self.phi.get_variables()), ['x1', 'x2', 'x3'])
# self.assertListEqual(sorted(self.phi1.get_variables()), ['x1', 'x2', 'x3'])
# def test_get_value(self):
# self.assertAlmostEqual(self.phi1.get_value({'x1': 1, 'x2': 2, 'x3': 0, 'x4': 2, 'x5': 1}), 5.0)
# self.assertAlmostEqual(self.phi1.get_value([1, 1, 1]), 9.0)
def test_get_cardinality(self):
self.assertEqual(self.phi.get_cardinality('x1'), 2)
self.assertEqual(self.phi.get_cardinality('x2'), 2)
self.assertEqual(self.phi.get_cardinality('x3'), 2)
def test_get_cardinality_scopeerror(self):
self.assertRaises(exceptions.ScopeError, self.phi.get_cardinality, 'x4')
def test_marginalize(self):
self.phi1.marginalize('x1')
np_test.assert_array_equal(self.phi1.values, np.array([6, 8, 10, 12, 14, 16]))
self.phi1.marginalize(['x2'])
np_test.assert_array_equal(self.phi1.values, np.array([30, 36]))
self.phi1.marginalize('x3')
np_test.assert_array_equal(self.phi1.values, np.array([66]))
def test_marginalize_scopeerror(self):
self.assertRaises(exceptions.ScopeError, self.phi.marginalize, 'x4')
self.assertRaises(exceptions.ScopeError, self.phi.marginalize, ['x4'])
self.phi.marginalize('x1')
self.assertRaises(exceptions.ScopeError, self.phi.marginalize, 'x1')
def test_normalize(self):
self.phi1.normalize()
np_test.assert_almost_equal(self.phi1.values, np.array(
[0, 0.01515152, 0.03030303, 0.04545455, 0.06060606,
0.07575758, 0.09090909, 0.10606061, 0.12121212,
0.13636364, 0.15151515, 0.16666667]))
# def test_maximize(self):
# self.phi1.maximize(['x1'])
# max_val_indices = {'x1':1}
# data = [max_val_indices] * 6
# #print(self.phi1)
# np_test.assert_array_equal(self.phi1.values, np.array([6,7,8,9,10,11]))
# self.assertEqual(self.phi1.data, sorted(data, key=lambda t: sorted(t.keys())))
def test_reduce(self):
self.phi1.reduce(['x1_0', 'x2_0'])
np_test.assert_array_equal(self.phi1.values, np.array([0, 1]))
def test_reduce_typeerror(self):
self.assertRaises(TypeError, self.phi1.reduce, 'x10')
self.assertRaises(TypeError, self.phi1.reduce, ['x10'])
def test_reduce_scopeerror(self):
self.assertRaises(exceptions.ScopeError, self.phi1.reduce, 'x4_1')
def test_reduce_sizeerror(self):
self.assertRaises(exceptions.SizeError, self.phi1.reduce, 'x3_5')
def test_identity_factor(self):
identity_factor = self.phi.identity_factor()
self.assertEquals(list(identity_factor.variables), ['x1', 'x2', 'x3'])
np_test.assert_array_equal(identity_factor.cardinality, [2, 2, 2])
np_test.assert_array_equal(identity_factor.values, np.ones(8))
def test__str(self):
from pgmpy import factors
phi = factors.Factor(['x1'], [2], np.ones(2))
self.assertEqual(repr(phi.__str__()), "'x1\\t\\tphi(x1)\\n------------------------\\nx1_0\\t\\t1.0\\nx1_1\\t\\t1.0\\n'")
def test_factor_product(self):
from pgmpy import factors
phi = factors.Factor(['x1', 'x2'], [2, 2], range(4))
phi1 = factors.Factor(['x3', 'x4'], [2, 2], range(4))
factor_product = factors.factor_product(phi, phi1)
np_test.assert_array_equal(factor_product.values,
np.array([0, 0, 0, 0, 0, 1,
2, 3, 0, 2, 4, 6,
0, 3, 6, 9]))
self.assertEqual(factor_product.variables, OrderedDict([
('x1', ['x1_0', 'x1_1']),
('x2', ['x2_0', 'x2_1']),
('x3', ['x3_0', 'x3_1']),
('x4', ['x4_0', 'x4_1'])]
))
phi = factors.Factor(['x1', 'x2'], [3, 2], range(6))
phi1 = factors.Factor(['x2', 'x3'], [2, 2], range(4))
factor_product = factors.factor_product(phi, phi1)
np_test.assert_array_equal(factor_product.values,
np.array([0, 0, 2, 3, 0, 2, 6, 9, 0, 4, 10, 15]))
self.assertEqual(factor_product.variables, OrderedDict(
[('x1', ['x1_0', 'x1_1', 'x1_2']),
('x2', ['x2_0', 'x2_1']),
('x3', ['x3_0', 'x3_1'])]))
def test_factor_product2(self):
from pgmpy import factors
phi = factors.Factor(['x1', 'x2'], [2, 2], range(4))
phi1 = factors.Factor(['x3', 'x4'], [2, 2], range(4))
factor_product = phi.product(phi1)
np_test.assert_array_equal(factor_product.values,
np.array([0, 0, 0, 0, 0, 1,
2, 3, 0, 2, 4, 6,
0, 3, 6, 9]))
self.assertEqual(factor_product.variables, OrderedDict([
('x1', ['x1_0', 'x1_1']),
('x2', ['x2_0', 'x2_1']),
('x3', ['x3_0', 'x3_1']),
('x4', ['x4_0', 'x4_1'])]
))
phi = factors.Factor(['x1', 'x2'], [3, 2], range(6))
phi1 = factors.Factor(['x2', 'x3'], [2, 2], range(4))
factor_product = phi.product(phi1)
np_test.assert_array_equal(factor_product.values,
np.array([0, 0, 2, 3, 0, 2, 6, 9, 0, 4, 10, 15]))
self.assertEqual(factor_product.variables, OrderedDict(
[('x1', ['x1_0', 'x1_1', 'x1_2']),
('x2', ['x2_0', 'x2_1']),
('x3', ['x3_0', 'x3_1'])]))
def test_factor_product_non_factor_arg(self):
from pgmpy import factors
self.assertRaises(TypeError, factors.factor_product, 1, 2)
def test_factor_mul(self):
from pgmpy import factors
phi = factors.Factor(['x1', 'x2'], [2, 2], range(4))
phi1 = factors.Factor(['x3', 'x4'], [2, 2], range(4))
factor_product = phi * phi1
np_test.assert_array_equal(factor_product.values,
np.array([0, 0, 0, 0, 0, 1,
2, 3, 0, 2, 4, 6,
0, 3, 6, 9]))
self.assertEqual(factor_product.variables, OrderedDict([
('x1', ['x1_0', 'x1_1']),
('x2', ['x2_0', 'x2_1']),
('x3', ['x3_0', 'x3_1']),
('x4', ['x4_0', 'x4_1'])]
))
def test_factor_divide(self):
from pgmpy import factors
phi1 = factors.Factor(['x1', 'x2'], [2, 2], [1, 2, 2, 4])
phi2 = factors.Factor(['x1'], [2], [1, 2])
factor_divide = phi1.divide(phi2)
phi3 = factors.Factor(['x1', 'x2'], [2, 2], [1, 2, 1, 2])
self.assertEqual(phi3, factor_divide)
def test_factor_divide_invalid(self):
from pgmpy import factors
phi1 = factors.Factor(['x1', 'x2', ], [2, 2], [1, 2, 3, 4])
phi2 = factors.Factor(['x1'], [2], [0, 2])
factor_divide = phi1.divide(phi2)
np_test.assert_array_equal(factor_divide.values, np.array([0, 0, 1.5, 2]))
def test_factor_divide_no_common_scope(self):
from pgmpy import factors
phi1 = factors.Factor(['x1', 'x2', ], [2, 2], [1, 2, 3, 4])
phi2 = factors.Factor(['x3'], [2], [0, 2])
self.assertRaises(ValueError, factors.factor_divide, phi1, phi2)
def test_factor_divide_non_factor_arg(self):
from pgmpy import factors
self.assertRaises(TypeError, factors.factor_divide, 1, 1)
# def test_sum_values(self):
# self.assertEqual(self.phi1.sum_values(), 66)
def test_eq(self):
self.assertFalse(self.phi == self.phi1)
self.assertTrue(self.phi == self.phi)
self.assertTrue(self.phi1 == self.phi1)
def tearDown(self):
del self.phi
del self.phi1
def setup(self):
self.phi = Factor(['x1', 'x2', 'x3'], [2, 2, 2], np.ones(8))
self.phi_large = Factor(range(10), [2] * 10, [1] * (2**10))
class TestFactorMethods(unittest.TestCase):
def setUp(self):
self.phi = Factor(['x1', 'x2', 'x3'], [2, 2, 2], np.random.uniform(5, 10, size=8))
self.phi1 = Factor(['x1', 'x2', 'x3'], [2, 3, 2], range(12))
def test_scope(self):
self.assertListEqual(self.phi.scope(), ['x1', 'x2', 'x3'])
self.assertListEqual(self.phi1.scope(), ['x1', 'x2', 'x3'])
def test_assignment(self):
self.assertListEqual(self.phi.assignment([0]), [[State('x1', 0), State('x2', 0), State('x3', 0)]])
self.assertListEqual(self.phi.assignment([4, 5, 6]), [[State('x1', 1), State('x2', 0), State('x3', 0)],
[State('x1', 1), State('x2', 0), State('x3', 1)],
[State('x1', 1), State('x2', 1), State('x3', 0)]])
self.assertListEqual(self.phi1.assignment(np.array([4, 5, 6])),
[[State('x1', 0), State('x2', 2), State('x3', 0)],
[State('x1', 0), State('x2', 2), State('x3', 1)],
[State('x1', 1), State('x2', 0), State('x3', 0)]])
def test_assignment_indexerror(self):
self.assertRaises(IndexError, self.phi.assignment, [10])
self.assertRaises(IndexError, self.phi.assignment, [1, 3, 10, 5])
self.assertRaises(IndexError, self.phi.assignment, np.array([1, 3, 10, 5]))
def test_get_cardinality(self):
self.assertEqual(self.phi.get_cardinality('x1'), 2)
self.assertEqual(self.phi.get_cardinality('x2'), 2)
self.assertEqual(self.phi.get_cardinality('x3'), 2)
def test_get_cardinality_scopeerror(self):
self.assertRaises(exceptions.ScopeError, self.phi.get_cardinality, 'x4')
def test_marginalize(self):
self.phi1.marginalize('x1')
np_test.assert_array_equal(self.phi1.values, np.array([6, 8, 10, 12, 14, 16]))
self.phi1.marginalize(['x2'])
np_test.assert_array_equal(self.phi1.values, np.array([30, 36]))
self.phi1.marginalize('x3')
np_test.assert_array_equal(self.phi1.values, np.array([66]))
def test_marginalize_scopeerror(self):
self.assertRaises(exceptions.ScopeError, self.phi.marginalize, 'x4')
self.assertRaises(exceptions.ScopeError, self.phi.marginalize, ['x4'])
self.phi.marginalize('x1')
self.assertRaises(exceptions.ScopeError, self.phi.marginalize, 'x1')
def test_normalize(self):
self.phi1.normalize()
np_test.assert_almost_equal(self.phi1.values, np.array(
[0, 0.01515152, 0.03030303, 0.04545455, 0.06060606,
0.07575758, 0.09090909, 0.10606061, 0.12121212,
0.13636364, 0.15151515, 0.16666667]))
def test_reduce(self):
self.phi1.reduce([('x1', 0), ('x2', 0)])
np_test.assert_array_equal(self.phi1.values, np.array([0, 1]))
def test_reduce1(self):
self.phi1.reduce([('x2', 0), ('x1', 0)])
np_test.assert_array_equal(self.phi1.values, np.array([0, 1]))
@unittest.skip
def test_complete_reduce(self):
self.phi1.reduce([('x1', 0), ('x2', 0), ('x3', 1)])
np_test.assert_array_equal(self.phi1.values, np.array([0]))
np_test.assert_array_equal(self.phi1.cardinality, np.array([]))
np_test.assert_array_equal(self.phi1.variables, OrderedDict())
def test_reduce_typeerror(self):
self.assertRaises(ValueError, self.phi1.reduce, 'x10')
self.assertRaises(ValueError, self.phi1.reduce, ['x10'])
def test_reduce_scopeerror(self):
self.assertRaises(ValueError, self.phi1.reduce, ('x4', 1))
def test_reduce_sizeerror(self):
self.assertRaises(IndexError, self.phi1.reduce, ('x3', 5))
def test_identity_factor(self):
identity_factor = self.phi.identity_factor()
self.assertEquals(list(identity_factor.variables), ['x1', 'x2', 'x3'])
np_test.assert_array_equal(identity_factor.cardinality, [2, 2, 2])
np_test.assert_array_equal(identity_factor.values, np.ones(8))
def test_factor_product(self):
phi = Factor(['x1', 'x2'], [2, 2], range(4))
phi1 = Factor(['x3', 'x4'], [2, 2], range(4))
prod = factor_product(phi, phi1)
np_test.assert_array_equal(prod.values,
np.array([0, 0, 0, 0, 0, 1,
2, 3, 0, 2, 4, 6,
0, 3, 6, 9]))
self.assertEqual(prod.variables, OrderedDict([
('x1', [State('x1', 0), State('x1', 1)]),
('x2', [State('x2', 0), State('x2', 1)]),
('x3', [State('x3', 0), State('x3', 1)]),
('x4', [State('x4', 0), State('x4', 1)])]
))
phi = Factor(['x1', 'x2'], [3, 2], range(6))
phi1 = Factor(['x2', 'x3'], [2, 2], range(4))
prod = factor_product(phi, phi1)
np_test.assert_array_equal(prod.values,
np.array([0, 0, 2, 3, 0, 2, 6, 9, 0, 4, 10, 15]))
self.assertEqual(prod.variables, OrderedDict(
[('x1', [State('x1', 0), State('x1', 1), State('x1', 2)]),
('x2', [State('x2', 0), State('x2', 1)]),
('x3', [State('x3', 0), State('x3', 1)])]))
def test_factor_product2(self):
from pgmpy import factors
phi = factors.Factor(['x1', 'x2'], [2, 2], range(4))
phi1 = factors.Factor(['x3', 'x4'], [2, 2], range(4))
prod = phi.product(phi1)
np_test.assert_array_equal(prod.values,
np.array([0, 0, 0, 0, 0, 1,
2, 3, 0, 2, 4, 6,
0, 3, 6, 9]))
self.assertEqual(prod.variables, OrderedDict([
('x1', [State('x1', 0), State('x1', 1)]),
('x2', [State('x2', 0), State('x2', 1)]),
('x3', [State('x3', 0), State('x3', 1)]),
('x4', [State('x4', 0), State('x4', 1)])]
))
phi = Factor(['x1', 'x2'], [3, 2], range(6))
phi1 = Factor(['x2', 'x3'], [2, 2], range(4))
prod = phi.product(phi1)
np_test.assert_array_equal(prod.values,
np.array([0, 0, 2, 3, 0, 2, 6, 9, 0, 4, 10, 15]))
self.assertEqual(prod.variables, OrderedDict(
[('x1', [State('x1', 0), State('x1', 1), State('x1', 2)]),
('x2', [State('x2', 0), State('x2', 1)]),
('x3', [State('x3', 0), State('x3', 1)])]))
def test_factor_product_non_factor_arg(self):
self.assertRaises(TypeError, factor_product, 1, 2)
def test_factor_mul(self):
phi = Factor(['x1', 'x2'], [2, 2], range(4))
phi1 = Factor(['x3', 'x4'], [2, 2], range(4))
prod = phi * phi1
np_test.assert_array_equal(prod.values,
np.array([0, 0, 0, 0, 0, 1,
2, 3, 0, 2, 4, 6,
0, 3, 6, 9]))
self.assertEqual(prod.variables, OrderedDict([
('x1', [State('x1', 0), State('x1', 1)]),
('x2', [State('x2', 0), State('x2', 1)]),
('x3', [State('x3', 0), State('x3', 1)]),
('x4', [State('x4', 0), State('x4', 1)])]
))
def test_factor_divide(self):
phi1 = Factor(['x1', 'x2'], [2, 2], [1, 2, 2, 4])
phi2 = Factor(['x1'], [2], [1, 2])
div = phi1.divide(phi2)
phi3 = Factor(['x1', 'x2'], [2, 2], [1, 2, 1, 2])
self.assertEqual(phi3, div)
def test_factor_divide_truediv(self):
phi1 = Factor(['x1', 'x2'], [2, 2], [1, 2, 2, 4])
phi2 = Factor(['x1'], [2], [1, 2])
div = phi1 / phi2
phi3 = Factor(['x1', 'x2'], [2, 2], [1, 2, 1, 2])
self.assertEqual(phi3, div)
# def test_factor_divide_dividebyzero(self):
# phi1 = Factor(['x1', 'x2'], [2, 2], [1, 2, 3, 4])
# phi2 = Factor(['x1'], [2], [0, 2])
# self.assertRaises(FloatingPointError, factor_divide, phi1, phi2)
#
def test_factor_divide_invalidvalue(self):
phi1 = Factor(['x1', 'x2'], [3, 2], [0.5, 0.2, 0, 0, 0.3, 0.45])
phi2 = Factor(['x1'], [3], [0.8, 0, 0.6])
div = phi1.divide(phi2)
np_test.assert_array_equal(div.values, np.array([0.625, 0.25, 0, 0, 0.5, 0.75]))
def test_factor_divide_no_common_scope(self):
phi1 = Factor(['x1', 'x2'], [2, 2], [1, 2, 3, 4])
phi2 = Factor(['x3'], [2], [0, 2])
self.assertRaises(ValueError, factor_divide, phi1, phi2)
def test_factor_divide_non_factor_arg(self):
self.assertRaises(TypeError, factor_divide, 1, 1)
def test_eq(self):
self.assertFalse(self.phi == self.phi1)
self.assertTrue(self.phi == self.phi)
self.assertTrue(self.phi1 == self.phi1)
def test_eq1(self):
phi1 = Factor(['x1', 'x2', 'x3'], [2, 4, 3], range(24))
phi2 = Factor(['x2', 'x1', 'x3'], [4, 2, 3], [0, 1, 2, 12, 13, 14, 3,
4, 5, 15, 16, 17, 6, 7,
8, 18, 19, 20, 9, 10, 11,
21, 22, 23])
self.assertTrue(phi1, phi2)
def test_index_for_assignment(self):
for i, j in enumerate(itertools.product(*[range(2), range(3), range(2)])):
self.assertEqual(self.phi1._index_for_assignment(j), i)
def test_maximize1(self):
self.phi1.maximize('x1')
self.assertEqual(self.phi1, Factor(['x2', 'x3'], [3, 2], [6, 7, 8, 9, 10, 11]))
self.phi1.maximize('x2')
self.assertEqual(self.phi1, Factor(['x3'], [2], [10, 11]))
def test_maximize2(self):
self.phi1.maximize(['x1', 'x2'])
self.assertEqual(self.phi1, Factor(['x3'], [2], [10, 11]))
def test_maximize3(self):
self.phi2 = Factor(['x1', 'x2', 'x3'], [3, 2, 2], [0.25, 0.35, 0.08, 0.16, 0.05, 0.07,
0.00, 0.00, 0.15, 0.21, 0.08, 0.18])
self.phi2.maximize('x2')
self.assertEqual(self.phi2, Factor(['x1', 'x3'], [3, 2], [0.25, 0.35, 0.05,
0.07, 0.15, 0.21]))
def tearDown(self):
del self.phi
del self.phi1
def test_factor_divide_invalidvalue(self):
phi1 = Factor(['x1', 'x2'], [3, 2], [0.5, 0.2, 0, 0, 0.3, 0.45])
phi2 = Factor(['x1'], [3], [0.8, 0, 0.6])
div = phi1.divide(phi2)
np_test.assert_array_equal(div.values, np.array([0.625, 0.25, 0, 0, 0.5, 0.75]))
def setUp(self):
self.phi1 = Factor(['x1', 'x2', 'x3'], [2, 3, 2], range(12))
self.phi2 = Factor(['x3', 'x4', 'x1'], [2, 2, 2], range(8))
self.phi3 = Factor(['x5', 'x6', 'x7'], [2, 2, 2], range(8))
self.phi4 = Factor(['x5', 'x7', 'x8'], [2, 2, 2], range(8))
def setUp(self):
self.phi = Factor(['x1', 'x2', 'x3'], [2, 2, 2], np.random.uniform(5, 10, size=8))
self.phi1 = Factor(['x1', 'x2', 'x3'], [2, 3, 2], range(12))
def backward_inference(self, variables, evidence=None):
"""
Backward inference method using belief propagation.
Parameters:
----------
variables: list
list of variables for which you want to compute the probability
evidence: dict
a dict key, value pair as {var: state_of_var_observed}
None if no evidence
Examples:
--------
>>> from pgmpy.factors import TabularCPD
>>> from pgmpy.models import DynamicBayesianNetwork as DBN
>>> from pgmpy.inference import DBNInference
>>> dbnet = DBN()
>>> dbnet.add_edges_from([(('Z', 0), ('X', 0)), (('X', 0), ('Y', 0)),
... (('Z', 0), ('Z', 1))])
>>> z_start_cpd = TabularCPD(('Z', 0), 2, [[0.5, 0.5]])
>>> x_i_cpd = TabularCPD(('X', 0), 2, [[0.6, 0.9],
... [0.4, 0.1]],
... evidence=[('Z', 0)],
... evidence_card=2)
>>> y_i_cpd = TabularCPD(('Y', 0), 2, [[0.2, 0.3],
... [0.8, 0.7]],
... evidence=[('X', 0)],
... evidence_card=2)
>>> z_trans_cpd = TabularCPD(('Z', 1), 2, [[0.4, 0.7],
... [0.6, 0.3]],
... evidence=[('Z', 0)],
... evidence_card=2)
>>> dbnet.add_cpds(z_start_cpd, z_trans_cpd, x_i_cpd, y_i_cpd)
>>> dbnet.initialize_initial_state()
>>> dbn_inf = DBNInference(dbnet)
>>> dbn_inf.backward_inference([('X', 0)], {('Y', 0):0, ('Y', 1):1, ('Y', 2):1})[('X', 0)].values
array([ 0.66594382, 0.33405618])
"""
variable_dict = defaultdict(list)
for var in variables:
variable_dict[var[1]].append(var)
time_range = max(variable_dict)
interface_nodes_dict = {}
if evidence:
evid_time_range = max(
[time_slice for var, time_slice in evidence.keys()])
time_range = max(time_range, evid_time_range)
end_bp = BeliefPropagation(self.start_junction_tree)
potential_dict = self.forward_inference(variables, evidence,
'potential')
update_factor = self._shift_factor(potential_dict[time_range], 1)
factor_values = {}
for time_slice in range(time_range, 0, -1):
evidence_time = self._get_evidence(evidence, time_slice, 1)
evidence_prev_time = self._get_evidence(evidence, time_slice - 1,
0)
if evidence_prev_time:
interface_nodes_dict = {
k: v
for k, v in evidence_prev_time.items()
if k in self.interface_nodes_0
}
if evidence_time:
evidence_time.update(interface_nodes_dict)
mid_bp = BeliefPropagation(self.one_and_half_junction_tree)
self._update_belief(mid_bp, self.in_clique,
potential_dict[time_slice - 1])
forward_factor = self._shift_factor(potential_dict[time_slice], 1)
self._update_belief(mid_bp, self.out_clique, forward_factor,
update_factor)
if variable_dict[time_slice]:
variable_time = self._shift_nodes(variable_dict[time_slice], 1)
new_values = mid_bp.query(variable_time,
evidence=evidence_time)
changed_values = {}
for key in new_values.keys():
new_key = (key[0], time_slice)
new_factor = Factor([new_key], new_values[key].cardinality,
new_values[key].values)
changed_values[new_key] = new_factor
factor_values.update(changed_values)
clique_phi = self._get_factor(mid_bp, evidence_time)
in_clique_phi = self._marginalize_factor(self.interface_nodes_0,
clique_phi)
update_factor = self._shift_factor(in_clique_phi, 1)
out_clique_phi = self._shift_factor(update_factor, 0)
self._update_belief(end_bp, self.start_interface_clique,
potential_dict[0], out_clique_phi)
evidence_0 = self._get_evidence(evidence, 0, 0)
if variable_dict[0]:
factor_values.update(end_bp.query(variable_dict[0], evidence_0))
return factor_values
