def main_brute(file, keep_atts, edges): perms = list(itertools.permutations(keep_atts)) max_widths, cliques, dictionaries, sat_runtimes, total_sat_runtime = find_widths_brute( file, keep_atts, edges) min_tree_width_idx = np.argmin(max_widths) min_tree_width = np.min(max_widths) best_ordering = perms[min_tree_width_idx] best_clique = cliques[min_tree_width_idx] best_dictionary = dictionaries[min_tree_width_idx] graph = create_bayes_net(file, keep_atts, edges) mgraph = graph.to_markov_model() inference = VariableElimination(mgraph) # Perform inference using the best ordering igraph = inference.induced_graph(list(best_ordering)) graph_size = len(keep_atts) density = nx.density(igraph) n_edges = igraph.number_of_edges() worst_runtime = np.max(sat_runtimes) print("Graph size ", graph_size) print("Graph density ", density) print("Graph edges ", n_edges) print("The minimum tree width is ", min_tree_width) print("The best ordering is ", best_ordering) print("The worst case SAT runtime is ", worst_runtime) print("The total SAT runtime is ", total_sat_runtime) return graph_size, density, n_edges, worst_runtime, total_sat_runtime
def find_widths(file, keep_atts, edges): graph = create_bayes_net(file, keep_atts, edges) N = len(keep_atts) perms = list(itertools.permutations(keep_atts)) ATT_DICT = {keep_atts[i]: i + 1 for i in range(0, len(keep_atts))} mgraph = graph.to_markov_model() inference = VariableElimination(mgraph) max_widths = [] cliques = [] dictionaries = [] sat_runtimes = [] total_sat_runtime = 0 for perm in tqdm.tqdm(perms): adj_matrix = np.zeros((N, N)) igraph = inference.induced_graph(list(perm)) edges = igraph.edges for ed in edges: first, second = ed first = ATT_DICT[first] - 1 second = ATT_DICT[second] - 1 adj_matrix[first, second] = 1 adj_matrix[second, first] = 1 adj_matrix[first, first] = 1 adj_matrix[second, second] = 1 adj_matrix = adj_matrix.astype(bool) max_width = 0 curr_clique = [] curr_dict = {} runtime = 0 start_time = time.time() for width in range(3, N): satisfiable, clique, diction = run_sat(width, adj_matrix, keep_atts) if satisfiable and width > max_width: max_width = width curr_clique = clique curr_dict = diction end_time = time.time() runtime = end_time - start_time total_sat_runtime += runtime max_widths.append(max_width) cliques.append(curr_clique) dictionaries.append(curr_dict) sat_runtimes.append(runtime) return np.array( max_widths), cliques, dictionaries, sat_runtimes, total_sat_runtime
class TestVariableElimination(unittest.TestCase): def setUp(self): self.bayesian_model = BayesianModel([('A', 'J'), ('R', 'J'), ('J', 'Q'), ('J', 'L'), ('G', 'L')]) cpd_a = TabularCPD('A', 2, [[0.2], [0.8]]) cpd_r = TabularCPD('R', 2, [[0.4], [0.6]]) cpd_j = TabularCPD('J', 2, [[0.9, 0.6, 0.7, 0.1], [0.1, 0.4, 0.3, 0.9]], ['R', 'A'], [2, 2]) cpd_q = TabularCPD('Q', 2, [[0.9, 0.2], [0.1, 0.8]], ['J'], [2]) cpd_l = TabularCPD('L', 2, [[0.9, 0.45, 0.8, 0.1], [0.1, 0.55, 0.2, 0.9]], ['G', 'J'], [2, 2]) cpd_g = TabularCPD('G', 2, [[0.6], [0.4]]) self.bayesian_model.add_cpds(cpd_a, cpd_g, cpd_j, cpd_l, cpd_q, cpd_r) self.bayesian_inference = VariableElimination(self.bayesian_model) # All the values that are used for comparision in the all the tests are # found using SAMIAM (assuming that it is correct ;)) def test_query_single_variable(self): query_result = self.bayesian_inference.query(['J']) np_test.assert_array_almost_equal(query_result['J'].values, np.array([0.416, 0.584])) def test_query_multiple_variable(self): query_result = self.bayesian_inference.query(['Q', 'J']) np_test.assert_array_almost_equal(query_result['J'].values, np.array([0.416, 0.584])) np_test.assert_array_almost_equal(query_result['Q'].values, np.array([0.4912, 0.5088])) def test_query_single_variable_with_evidence(self): query_result = self.bayesian_inference.query(variables=['J'], evidence={ 'A': 0, 'R': 1 }) np_test.assert_array_almost_equal(query_result['J'].values, np.array([0.60, 0.40])) def test_query_multiple_variable_with_evidence(self): query_result = self.bayesian_inference.query(variables=['J', 'Q'], evidence={ 'A': 0, 'R': 0, 'G': 0, 'L': 1 }) np_test.assert_array_almost_equal(query_result['J'].values, np.array([0.818182, 0.181818])) np_test.assert_array_almost_equal(query_result['Q'].values, np.array([0.772727, 0.227273])) def test_query_multiple_times(self): # This just tests that the models are not getting modified while querying them query_result = self.bayesian_inference.query(['J']) query_result = self.bayesian_inference.query(['J']) np_test.assert_array_almost_equal(query_result['J'].values, np.array([0.416, 0.584])) query_result = self.bayesian_inference.query(['Q', 'J']) query_result = self.bayesian_inference.query(['Q', 'J']) np_test.assert_array_almost_equal(query_result['J'].values, np.array([0.416, 0.584])) np_test.assert_array_almost_equal(query_result['Q'].values, np.array([0.4912, 0.5088])) query_result = self.bayesian_inference.query(variables=['J'], evidence={ 'A': 0, 'R': 1 }) query_result = self.bayesian_inference.query(variables=['J'], evidence={ 'A': 0, 'R': 1 }) np_test.assert_array_almost_equal(query_result['J'].values, np.array([0.60, 0.40])) query_result = self.bayesian_inference.query(variables=['J', 'Q'], evidence={ 'A': 0, 'R': 0, 'G': 0, 'L': 1 }) query_result = self.bayesian_inference.query(variables=['J', 'Q'], evidence={ 'A': 0, 'R': 0, 'G': 0, 'L': 1 }) np_test.assert_array_almost_equal(query_result['J'].values, np.array([0.818182, 0.181818])) np_test.assert_array_almost_equal(query_result['Q'].values, np.array([0.772727, 0.227273])) def test_max_marginal(self): np_test.assert_almost_equal(self.bayesian_inference.max_marginal(), 0.1659, decimal=4) def test_max_marginal_var(self): np_test.assert_almost_equal(self.bayesian_inference.max_marginal(['G' ]), 0.5714, decimal=4) def test_max_marginal_var1(self): np_test.assert_almost_equal(self.bayesian_inference.max_marginal( ['G', 'R']), 0.4055, decimal=4) def test_max_marginal_var2(self): np_test.assert_almost_equal(self.bayesian_inference.max_marginal( ['G', 'R', 'A']), 0.3260, decimal=4) def test_map_query(self): map_query = self.bayesian_inference.map_query() self.assertDictEqual(map_query, { 'A': 1, 'R': 1, 'J': 1, 'Q': 1, 'G': 0, 'L': 0 }) def test_map_query_with_evidence(self): map_query = self.bayesian_inference.map_query(['A', 'R', 'L'], { 'J': 0, 'Q': 1, 'G': 0 }) self.assertDictEqual(map_query, {'A': 1, 'R': 0, 'L': 0}) def test_induced_graph(self): induced_graph = self.bayesian_inference.induced_graph( ['G', 'Q', 'A', 'J', 'L', 'R']) result_edges = sorted([sorted(x) for x in induced_graph.edges()]) self.assertEqual([['A', 'J'], ['A', 'R'], ['G', 'J'], ['G', 'L'], ['J', 'L'], ['J', 'Q'], ['J', 'R'], ['L', 'R']], result_edges) def test_induced_width(self): result_width = self.bayesian_inference.induced_width( ['G', 'Q', 'A', 'J', 'L', 'R']) self.assertEqual(2, result_width) def tearDown(self): del self.bayesian_inference del self.bayesian_model
class TestVariableElimination(unittest.TestCase): def setUp(self): self.bayesian_model = BayesianModel([("A", "J"), ("R", "J"), ("J", "Q"), ("J", "L"), ("G", "L")]) cpd_a = TabularCPD("A", 2, values=[[0.2], [0.8]]) cpd_r = TabularCPD("R", 2, values=[[0.4], [0.6]]) cpd_j = TabularCPD( "J", 2, values=[[0.9, 0.6, 0.7, 0.1], [0.1, 0.4, 0.3, 0.9]], evidence=["A", "R"], evidence_card=[2, 2], ) cpd_q = TabularCPD("Q", 2, values=[[0.9, 0.2], [0.1, 0.8]], evidence=["J"], evidence_card=[2]) cpd_l = TabularCPD( "L", 2, values=[[0.9, 0.45, 0.8, 0.1], [0.1, 0.55, 0.2, 0.9]], evidence=["J", "G"], evidence_card=[2, 2], ) cpd_g = TabularCPD("G", 2, values=[[0.6], [0.4]]) self.bayesian_model.add_cpds(cpd_a, cpd_g, cpd_j, cpd_l, cpd_q, cpd_r) self.bayesian_inference = VariableElimination(self.bayesian_model) # All the values that are used for comparision in the all the tests are # found using SAMIAM (assuming that it is correct ;)) def test_query_single_variable(self): query_result = self.bayesian_inference.query(["J"]) self.assertEqual( query_result, DiscreteFactor(variables=["J"], cardinality=[2], values=[0.416, 0.584]), ) def test_query_multiple_variable(self): query_result = self.bayesian_inference.query(["Q", "J"]) self.assertEqual( query_result, DiscreteFactor( variables=["J", "Q"], cardinality=[2, 2], values=np.array([[0.3744, 0.0416], [0.1168, 0.4672]]), ), ) def test_query_single_variable_with_evidence(self): query_result = self.bayesian_inference.query(variables=["J"], evidence={ "A": 0, "R": 1 }) self.assertEqual( query_result, DiscreteFactor(variables=["J"], cardinality=[2], values=[0.6, 0.4]), ) def test_query_multiple_variable_with_evidence(self): query_result = self.bayesian_inference.query(variables=["J", "Q"], evidence={ "A": 0, "R": 0, "G": 0, "L": 1 }) self.assertEqual( query_result, DiscreteFactor( variables=["J", "Q"], cardinality=[2, 2], values=np.array([[0.73636364, 0.08181818], [0.03636364, 0.14545455]]), ), ) def test_query_multiple_times(self): # This just tests that the models are not getting modified while querying them query_result = self.bayesian_inference.query(["J"]) query_result = self.bayesian_inference.query(["J"]) self.assertEqual( query_result, DiscreteFactor(variables=["J"], cardinality=[2], values=np.array([0.416, 0.584])), ) query_result = self.bayesian_inference.query(["Q", "J"]) query_result = self.bayesian_inference.query(["Q", "J"]) self.assertEqual( query_result, DiscreteFactor( variables=["J", "Q"], cardinality=[2, 2], values=np.array([[0.3744, 0.0416], [0.1168, 0.4672]]), ), ) query_result = self.bayesian_inference.query(variables=["J"], evidence={ "A": 0, "R": 1 }) query_result = self.bayesian_inference.query(variables=["J"], evidence={ "A": 0, "R": 1 }) self.assertEqual( query_result, DiscreteFactor(variables=["J"], cardinality=[2], values=[0.6, 0.4]), ) query_result = self.bayesian_inference.query(variables=["J", "Q"], evidence={ "A": 0, "R": 0, "G": 0, "L": 1 }) query_result = self.bayesian_inference.query(variables=["J", "Q"], evidence={ "A": 0, "R": 0, "G": 0, "L": 1 }) self.assertEqual( query_result, DiscreteFactor( variables=["J", "Q"], cardinality=[2, 2], values=np.array([[0.73636364, 0.08181818], [0.03636364, 0.14545455]]), ), ) def test_max_marginal(self): np_test.assert_almost_equal(self.bayesian_inference.max_marginal(), 0.1659, decimal=4) def test_max_marginal_var(self): np_test.assert_almost_equal(self.bayesian_inference.max_marginal(["G" ]), 0.5714, decimal=4) def test_max_marginal_var1(self): np_test.assert_almost_equal(self.bayesian_inference.max_marginal( ["G", "R"]), 0.3740, decimal=4) def test_max_marginal_var2(self): np_test.assert_almost_equal(self.bayesian_inference.max_marginal( ["G", "R", "A"]), 0.3061, decimal=4) def test_map_query(self): map_query = self.bayesian_inference.map_query() self.assertDictEqual(map_query, { "A": 1, "R": 1, "J": 1, "Q": 1, "G": 0, "L": 0 }) def test_map_query_with_evidence(self): map_query = self.bayesian_inference.map_query(["A", "R", "L"], { "J": 0, "Q": 1, "G": 0 }) self.assertDictEqual(map_query, {"A": 1, "R": 0, "L": 0}) def test_induced_graph(self): induced_graph = self.bayesian_inference.induced_graph( ["G", "Q", "A", "J", "L", "R"]) result_edges = sorted([sorted(x) for x in induced_graph.edges()]) self.assertEqual( [ ["A", "J"], ["A", "R"], ["G", "J"], ["G", "L"], ["J", "L"], ["J", "Q"], ["J", "R"], ["L", "R"], ], result_edges, ) def test_induced_width(self): result_width = self.bayesian_inference.induced_width( ["G", "Q", "A", "J", "L", "R"]) self.assertEqual(2, result_width) def tearDown(self): del self.bayesian_inference del self.bayesian_model
class TestVariableEliminationMarkov(unittest.TestCase): def setUp(self): # It is just a moralised version of the above Bayesian network so all the results are same. Only factors # are under consideration for inference so this should be fine. self.markov_model = MarkovModel([ ("A", "J"), ("R", "J"), ("J", "Q"), ("J", "L"), ("G", "L"), ("A", "R"), ("J", "G"), ]) factor_a = TabularCPD("A", 2, values=[[0.2], [0.8]]).to_factor() factor_r = TabularCPD("R", 2, values=[[0.4], [0.6]]).to_factor() factor_j = TabularCPD( "J", 2, values=[[0.9, 0.6, 0.7, 0.1], [0.1, 0.4, 0.3, 0.9]], evidence=["A", "R"], evidence_card=[2, 2], ).to_factor() factor_q = TabularCPD("Q", 2, values=[[0.9, 0.2], [0.1, 0.8]], evidence=["J"], evidence_card=[2]).to_factor() factor_l = TabularCPD( "L", 2, values=[[0.9, 0.45, 0.8, 0.1], [0.1, 0.55, 0.2, 0.9]], evidence=["J", "G"], evidence_card=[2, 2], ).to_factor() factor_g = TabularCPD("G", 2, [[0.6], [0.4]]).to_factor() self.markov_model.add_factors(factor_a, factor_r, factor_j, factor_q, factor_l, factor_g) self.markov_inference = VariableElimination(self.markov_model) # All the values that are used for comparision in the all the tests are # found using SAMIAM (assuming that it is correct ;)) def test_query_single_variable(self): query_result = self.markov_inference.query(["J"]) self.assertEqual( query_result, DiscreteFactor(variables=["J"], cardinality=[2], values=np.array([0.416, 0.584])), ) def test_query_multiple_variable(self): query_result = self.markov_inference.query(["Q", "J"]) self.assertEqual( query_result, DiscreteFactor( variables=["Q", "J"], cardinality=[2, 2], values=np.array([[0.3744, 0.1168], [0.0416, 0.4672]]), ), ) def test_query_single_variable_with_evidence(self): query_result = self.markov_inference.query(variables=["J"], evidence={ "A": 0, "R": 1 }) self.assertEqual( query_result, DiscreteFactor(variables=["J"], cardinality=[2], values=[0.6, 0.4]), ) def test_query_multiple_variable_with_evidence(self): query_result = self.markov_inference.query(variables=["J", "Q"], evidence={ "A": 0, "R": 0, "G": 0, "L": 1 }) self.assertEqual( query_result, DiscreteFactor( variables=["Q", "J"], cardinality=[2, 2], values=np.array([[0.081, 0.004], [0.009, 0.016]]), ), ) def test_query_multiple_times(self): # This just tests that the models are not getting modified while querying them query_result = self.markov_inference.query(["J"]) query_result = self.markov_inference.query(["J"]) self.assertEqual( query_result, DiscreteFactor(variables=["J"], cardinality=[2], values=np.array([0.416, 0.584])), ) query_result = self.markov_inference.query(["Q", "J"]) query_result = self.markov_inference.query(["Q", "J"]) self.assertEqual( query_result, DiscreteFactor( variables=["Q", "J"], cardinality=[2, 2], values=np.array([[0.3744, 0.1168], [0.0416, 0.4672]]), ), ) query_result = self.markov_inference.query(variables=["J"], evidence={ "A": 0, "R": 1 }) query_result = self.markov_inference.query(variables=["J"], evidence={ "A": 0, "R": 1 }) self.assertEqual( query_result, DiscreteFactor(variables=["J"], cardinality=[2], values=[0.6, 0.4]), ) query_result = self.markov_inference.query(variables=["J", "Q"], evidence={ "A": 0, "R": 0, "G": 0, "L": 1 }) query_result = self.markov_inference.query(variables=["J", "Q"], evidence={ "A": 0, "R": 0, "G": 0, "L": 1 }) self.assertEqual( query_result, DiscreteFactor( variables=["Q", "J"], cardinality=[2, 2], values=np.array([[0.081, 0.004], [0.009, 0.016]]), ), ) def test_max_marginal(self): np_test.assert_almost_equal(self.markov_inference.max_marginal(), 0.1659, decimal=4) def test_max_marginal_var(self): np_test.assert_almost_equal(self.markov_inference.max_marginal(["G"]), 0.1659, decimal=4) def test_max_marginal_var1(self): np_test.assert_almost_equal(self.markov_inference.max_marginal( ["G", "R"]), 0.1659, decimal=4) def test_max_marginal_var2(self): np_test.assert_almost_equal(self.markov_inference.max_marginal( ["G", "R", "A"]), 0.1659, decimal=4) def test_map_query(self): map_query = self.markov_inference.map_query() self.assertDictEqual(map_query, { "A": 1, "R": 1, "J": 1, "Q": 1, "G": 0, "L": 0 }) def test_map_query_with_evidence(self): map_query = self.markov_inference.map_query(["A", "R", "L"], { "J": 0, "Q": 1, "G": 0 }) self.assertDictEqual(map_query, {"A": 1, "R": 0, "L": 0}) def test_induced_graph(self): induced_graph = self.markov_inference.induced_graph( ["G", "Q", "A", "J", "L", "R"]) result_edges = sorted([sorted(x) for x in induced_graph.edges()]) self.assertEqual( [ ["A", "J"], ["A", "R"], ["G", "J"], ["G", "L"], ["J", "L"], ["J", "Q"], ["J", "R"], ["L", "R"], ], result_edges, ) def test_induced_width(self): result_width = self.markov_inference.induced_width( ["G", "Q", "A", "J", "L", "R"]) self.assertEqual(2, result_width) def tearDown(self): del self.markov_inference del self.markov_model
resurant_inference = VariableElimination(resurant) # Doing simple queries over one or multiple variables resurant_inference.query(variables=['location']) resurant_inference.query(variables=['location','no_of_people']) resurant_inference.query(variables=['no_of_people'], evidence={'location':1, 'quality':1}) # If we have evidence resurant_inference.query(variables=['no_of_people'], evidence={'location':1}, elimination_order=['quality', 'cost']) # can sepcify elimination sequence / otherwise system will choose automatically -2- " Induced Graph " " also defined as the undirected graph constructed by the unionof all the graphs formed in each step of variable elimination " # Check induced graph induced_graph = resurant_inference.induced_graph(['cost', 'location', 'no_of_people', 'quality']) induced_graph.nodes() " ['cost', 'location', 'no_of_people', 'quality'] " induced_graph.edges() " [('location', 'quality'), " " ('location', 'cost'), " " ('location ......... " " (...................)] " -3- " Belief Propagation - Information passage during elimination " ** " Clique Tree - also called 'Junction Tree' " " Undirected graph over a set of factors, where each factor represents a cluster of random variables and edges connect the clusters, " " where scope has a nonempty intersection. "
class TestVariableElimination(unittest.TestCase): def setUp(self): self.bayesian_model = BayesianModel([('A', 'J'), ('R', 'J'), ('J', 'Q'), ('J', 'L'), ('G', 'L')]) cpd_a = TabularCPD('A', 2, [[0.2], [0.8]]) cpd_r = TabularCPD('R', 2, [[0.4], [0.6]]) cpd_j = TabularCPD('J', 2, [[0.9, 0.6, 0.7, 0.1], [0.1, 0.4, 0.3, 0.9]], ['R', 'A'], [2, 2]) cpd_q = TabularCPD('Q', 2, [[0.9, 0.2], [0.1, 0.8]], ['J'], [2]) cpd_l = TabularCPD('L', 2, [[0.9, 0.45, 0.8, 0.1], [0.1, 0.55, 0.2, 0.9]], ['G', 'J'], [2, 2]) cpd_g = TabularCPD('G', 2, [[0.6], [0.4]]) self.bayesian_model.add_cpds(cpd_a, cpd_g, cpd_j, cpd_l, cpd_q, cpd_r) self.bayesian_inference = VariableElimination(self.bayesian_model) # All the values that are used for comparision in the all the tests are # found using SAMIAM (assuming that it is correct ;)) def test_query_single_variable(self): query_result = self.bayesian_inference.query(['J']) np_test.assert_array_almost_equal(query_result['J'].values, np.array([0.416, 0.584])) def test_query_multiple_variable(self): query_result = self.bayesian_inference.query(['Q', 'J']) np_test.assert_array_almost_equal(query_result['J'].values, np.array([0.416, 0.584])) np_test.assert_array_almost_equal(query_result['Q'].values, np.array([0.4912, 0.5088])) def test_query_single_variable_with_evidence(self): query_result = self.bayesian_inference.query(variables=['J'], evidence={'A': 0, 'R': 1}) np_test.assert_array_almost_equal(query_result['J'].values, np.array([0.60, 0.40])) def test_query_multiple_variable_with_evidence(self): query_result = self.bayesian_inference.query(variables=['J', 'Q'], evidence={'A': 0, 'R': 0, 'G': 0, 'L': 1}) np_test.assert_array_almost_equal(query_result['J'].values, np.array([0.818182, 0.181818])) np_test.assert_array_almost_equal(query_result['Q'].values, np.array([0.772727, 0.227273])) def test_query_multiple_times(self): # This just tests that the models are not getting modified while querying them query_result = self.bayesian_inference.query(['J']) query_result = self.bayesian_inference.query(['J']) np_test.assert_array_almost_equal(query_result['J'].values, np.array([0.416, 0.584])) query_result = self.bayesian_inference.query(['Q', 'J']) query_result = self.bayesian_inference.query(['Q', 'J']) np_test.assert_array_almost_equal(query_result['J'].values, np.array([0.416, 0.584])) np_test.assert_array_almost_equal(query_result['Q'].values, np.array([0.4912, 0.5088])) query_result = self.bayesian_inference.query(variables=['J'], evidence={'A': 0, 'R': 1}) query_result = self.bayesian_inference.query(variables=['J'], evidence={'A': 0, 'R': 1}) np_test.assert_array_almost_equal(query_result['J'].values, np.array([0.60, 0.40])) query_result = self.bayesian_inference.query(variables=['J', 'Q'], evidence={'A': 0, 'R': 0, 'G': 0, 'L': 1}) query_result = self.bayesian_inference.query(variables=['J', 'Q'], evidence={'A': 0, 'R': 0, 'G': 0, 'L': 1}) np_test.assert_array_almost_equal(query_result['J'].values, np.array([0.818182, 0.181818])) np_test.assert_array_almost_equal(query_result['Q'].values, np.array([0.772727, 0.227273])) def test_max_marginal(self): np_test.assert_almost_equal(self.bayesian_inference.max_marginal(), 0.1659, decimal=4) def test_max_marginal_var(self): np_test.assert_almost_equal(self.bayesian_inference.max_marginal(['G']), 0.5714, decimal=4) def test_max_marginal_var1(self): np_test.assert_almost_equal(self.bayesian_inference.max_marginal(['G', 'R']), 0.4055, decimal=4) def test_max_marginal_var2(self): np_test.assert_almost_equal(self.bayesian_inference.max_marginal(['G', 'R', 'A']), 0.3260, decimal=4) def test_map_query(self): map_query = self.bayesian_inference.map_query() self.assertDictEqual(map_query, {'A': 1, 'R': 1, 'J': 1, 'Q': 1, 'G': 0, 'L': 0}) def test_map_query_with_evidence(self): map_query = self.bayesian_inference.map_query(['A', 'R', 'L'], {'J': 0, 'Q': 1, 'G': 0}) self.assertDictEqual(map_query, {'A': 1, 'R': 0, 'L': 0}) def test_induced_graph(self): induced_graph = self.bayesian_inference.induced_graph(['G', 'Q', 'A', 'J', 'L', 'R']) result_edges = sorted([sorted(x) for x in induced_graph.edges()]) self.assertEqual([['A', 'J'], ['A', 'R'], ['G', 'J'], ['G', 'L'], ['J', 'L'], ['J', 'Q'], ['J', 'R'], ['L', 'R']], result_edges) def test_induced_width(self): result_width = self.bayesian_inference.induced_width(['G', 'Q', 'A', 'J', 'L', 'R']) self.assertEqual(2, result_width) def tearDown(self): del self.bayesian_inference del self.bayesian_model
class TestVariableEliminationMarkov(unittest.TestCase): def setUp(self): # It is just a moralised version of the above Bayesian network so all the results are same. Only factors # are under consideration for inference so this should be fine. self.markov_model = MarkovModel([('A', 'J'), ('R', 'J'), ('J', 'Q'), ('J', 'L'), ('G', 'L'), ('A', 'R'), ('J', 'G')]) factor_a = TabularCPD('A', 2, values=[[0.2], [0.8]]).to_factor() factor_r = TabularCPD('R', 2, values=[[0.4], [0.6]]).to_factor() factor_j = TabularCPD('J', 2, values=[[0.9, 0.6, 0.7, 0.1], [0.1, 0.4, 0.3, 0.9]], evidence=['A', 'R'], evidence_card=[2, 2]).to_factor() factor_q = TabularCPD('Q', 2, values=[[0.9, 0.2], [0.1, 0.8]], evidence=['J'], evidence_card=[2]).to_factor() factor_l = TabularCPD('L', 2, values=[[0.9, 0.45, 0.8, 0.1], [0.1, 0.55, 0.2, 0.9]], evidence=['J', 'G'], evidence_card=[2, 2]).to_factor() factor_g = TabularCPD('G', 2, [[0.6], [0.4]]).to_factor() self.markov_model.add_factors(factor_a, factor_r, factor_j, factor_q, factor_l, factor_g) self.markov_inference = VariableElimination(self.markov_model) # All the values that are used for comparision in the all the tests are # found using SAMIAM (assuming that it is correct ;)) def test_query_single_variable(self): query_result = self.markov_inference.query(['J']) np_test.assert_array_almost_equal(query_result['J'].values, np.array([0.416, 0.584])) def test_query_multiple_variable(self): query_result = self.markov_inference.query(['Q', 'J']) np_test.assert_array_almost_equal(query_result['J'].values, np.array([0.416, 0.584])) np_test.assert_array_almost_equal(query_result['Q'].values, np.array([0.4912, 0.5088])) def test_query_single_variable_with_evidence(self): query_result = self.markov_inference.query(variables=['J'], evidence={'A': 0, 'R': 1}) np_test.assert_array_almost_equal(query_result['J'].values, np.array([0.60, 0.40])) def test_query_multiple_variable_with_evidence(self): query_result = self.markov_inference.query(variables=['J', 'Q'], evidence={'A': 0, 'R': 0, 'G': 0, 'L': 1}) np_test.assert_array_almost_equal(query_result['J'].values, np.array([0.818182, 0.181818])) np_test.assert_array_almost_equal(query_result['Q'].values, np.array([0.772727, 0.227273])) def test_query_multiple_times(self): # This just tests that the models are not getting modified while querying them query_result = self.markov_inference.query(['J']) query_result = self.markov_inference.query(['J']) np_test.assert_array_almost_equal(query_result['J'].values, np.array([0.416, 0.584])) query_result = self.markov_inference.query(['Q', 'J']) query_result = self.markov_inference.query(['Q', 'J']) np_test.assert_array_almost_equal(query_result['J'].values, np.array([0.416, 0.584])) np_test.assert_array_almost_equal(query_result['Q'].values, np.array([0.4912, 0.5088])) query_result = self.markov_inference.query(variables=['J'], evidence={'A': 0, 'R': 1}) query_result = self.markov_inference.query(variables=['J'], evidence={'A': 0, 'R': 1}) np_test.assert_array_almost_equal(query_result['J'].values, np.array([0.60, 0.40])) query_result = self.markov_inference.query(variables=['J', 'Q'], evidence={'A': 0, 'R': 0, 'G': 0, 'L': 1}) query_result = self.markov_inference.query(variables=['J', 'Q'], evidence={'A': 0, 'R': 0, 'G': 0, 'L': 1}) np_test.assert_array_almost_equal(query_result['J'].values, np.array([0.818182, 0.181818])) np_test.assert_array_almost_equal(query_result['Q'].values, np.array([0.772727, 0.227273])) def test_max_marginal(self): np_test.assert_almost_equal(self.markov_inference.max_marginal(), 0.1659, decimal=4) def test_max_marginal_var(self): np_test.assert_almost_equal(self.markov_inference.max_marginal(['G']), 0.5714, decimal=4) def test_max_marginal_var1(self): np_test.assert_almost_equal(self.markov_inference.max_marginal(['G', 'R']), 0.4055, decimal=4) def test_max_marginal_var2(self): np_test.assert_almost_equal(self.markov_inference.max_marginal(['G', 'R', 'A']), 0.3260, decimal=4) def test_map_query(self): map_query = self.markov_inference.map_query() self.assertDictEqual(map_query, {'A': 1, 'R': 1, 'J': 1, 'Q': 1, 'G': 0, 'L': 0}) def test_map_query_with_evidence(self): map_query = self.markov_inference.map_query(['A', 'R', 'L'], {'J': 0, 'Q': 1, 'G': 0}) self.assertDictEqual(map_query, {'A': 1, 'R': 0, 'L': 0}) def test_induced_graph(self): induced_graph = self.markov_inference.induced_graph(['G', 'Q', 'A', 'J', 'L', 'R']) result_edges = sorted([sorted(x) for x in induced_graph.edges()]) self.assertEqual([['A', 'J'], ['A', 'R'], ['G', 'J'], ['G', 'L'], ['J', 'L'], ['J', 'Q'], ['J', 'R'], ['L', 'R']], result_edges) def test_induced_width(self): result_width = self.markov_inference.induced_width(['G', 'Q', 'A', 'J', 'L', 'R']) self.assertEqual(2, result_width) def tearDown(self): del self.markov_inference del self.markov_model
class TestVariableEliminationMarkov(unittest.TestCase): def setUp(self): # It is just a moralised version of the above Bayesian network so all the results are same. Only factors # are under consideration for inference so this should be fine. self.markov_model = MarkovModel([('A', 'J'), ('R', 'J'), ('J', 'Q'), ('J', 'L'), ('G', 'L'), ('A', 'R'), ('J', 'G')]) factor_a = TabularCPD('A', 2, values=[[0.2], [0.8]]).to_factor() factor_r = TabularCPD('R', 2, values=[[0.4], [0.6]]).to_factor() factor_j = TabularCPD('J', 2, values=[[0.9, 0.6, 0.7, 0.1], [0.1, 0.4, 0.3, 0.9]], evidence=['A', 'R'], evidence_card=[2, 2]).to_factor() factor_q = TabularCPD('Q', 2, values=[[0.9, 0.2], [0.1, 0.8]], evidence=['J'], evidence_card=[2]).to_factor() factor_l = TabularCPD('L', 2, values=[[0.9, 0.45, 0.8, 0.1], [0.1, 0.55, 0.2, 0.9]], evidence=['J', 'G'], evidence_card=[2, 2]).to_factor() factor_g = TabularCPD('G', 2, [[0.6], [0.4]]).to_factor() self.markov_model.add_factors(factor_a, factor_r, factor_j, factor_q, factor_l, factor_g) self.markov_inference = VariableElimination(self.markov_model) # All the values that are used for comparision in the all the tests are # found using SAMIAM (assuming that it is correct ;)) def test_query_single_variable(self): query_result = self.markov_inference.query(['J']) np_test.assert_array_almost_equal(query_result['J'].values, np.array([0.416, 0.584])) def test_query_multiple_variable(self): query_result = self.markov_inference.query(['Q', 'J']) np_test.assert_array_almost_equal(query_result['J'].values, np.array([0.416, 0.584])) np_test.assert_array_almost_equal(query_result['Q'].values, np.array([0.4912, 0.5088])) def test_query_single_variable_with_evidence(self): query_result = self.markov_inference.query(variables=['J'], evidence={ 'A': 0, 'R': 1 }) np_test.assert_array_almost_equal(query_result['J'].values, np.array([0.60, 0.40])) def test_query_multiple_variable_with_evidence(self): query_result = self.markov_inference.query(variables=['J', 'Q'], evidence={ 'A': 0, 'R': 0, 'G': 0, 'L': 1 }) np_test.assert_array_almost_equal(query_result['J'].values, np.array([0.818182, 0.181818])) np_test.assert_array_almost_equal(query_result['Q'].values, np.array([0.772727, 0.227273])) def test_query_multiple_times(self): # This just tests that the models are not getting modified while querying them query_result = self.markov_inference.query(['J']) query_result = self.markov_inference.query(['J']) np_test.assert_array_almost_equal(query_result['J'].values, np.array([0.416, 0.584])) query_result = self.markov_inference.query(['Q', 'J']) query_result = self.markov_inference.query(['Q', 'J']) np_test.assert_array_almost_equal(query_result['J'].values, np.array([0.416, 0.584])) np_test.assert_array_almost_equal(query_result['Q'].values, np.array([0.4912, 0.5088])) query_result = self.markov_inference.query(variables=['J'], evidence={ 'A': 0, 'R': 1 }) query_result = self.markov_inference.query(variables=['J'], evidence={ 'A': 0, 'R': 1 }) np_test.assert_array_almost_equal(query_result['J'].values, np.array([0.60, 0.40])) query_result = self.markov_inference.query(variables=['J', 'Q'], evidence={ 'A': 0, 'R': 0, 'G': 0, 'L': 1 }) query_result = self.markov_inference.query(variables=['J', 'Q'], evidence={ 'A': 0, 'R': 0, 'G': 0, 'L': 1 }) np_test.assert_array_almost_equal(query_result['J'].values, np.array([0.818182, 0.181818])) np_test.assert_array_almost_equal(query_result['Q'].values, np.array([0.772727, 0.227273])) def test_max_marginal(self): np_test.assert_almost_equal(self.markov_inference.max_marginal(), 0.1659, decimal=4) def test_max_marginal_var(self): np_test.assert_almost_equal(self.markov_inference.max_marginal(['G']), 0.5714, decimal=4) def test_max_marginal_var1(self): np_test.assert_almost_equal(self.markov_inference.max_marginal( ['G', 'R']), 0.4055, decimal=4) def test_max_marginal_var2(self): np_test.assert_almost_equal(self.markov_inference.max_marginal( ['G', 'R', 'A']), 0.3260, decimal=4) def test_map_query(self): map_query = self.markov_inference.map_query() self.assertDictEqual(map_query, { 'A': 1, 'R': 1, 'J': 1, 'Q': 1, 'G': 0, 'L': 0 }) def test_map_query_with_evidence(self): map_query = self.markov_inference.map_query(['A', 'R', 'L'], { 'J': 0, 'Q': 1, 'G': 0 }) self.assertDictEqual(map_query, {'A': 1, 'R': 0, 'L': 0}) def test_induced_graph(self): induced_graph = self.markov_inference.induced_graph( ['G', 'Q', 'A', 'J', 'L', 'R']) result_edges = sorted([sorted(x) for x in induced_graph.edges()]) self.assertEqual([['A', 'J'], ['A', 'R'], ['G', 'J'], ['G', 'L'], ['J', 'L'], ['J', 'Q'], ['J', 'R'], ['L', 'R']], result_edges) def test_induced_width(self): result_width = self.markov_inference.induced_width( ['G', 'Q', 'A', 'J', 'L', 'R']) self.assertEqual(2, result_width) def tearDown(self): del self.markov_inference del self.markov_model
a = table_values[a2][1:16] seen_dat.iloc[a1, 18:33] = a print(seen_dat) a2 = 0 break if i not in j: a2 = 0 seen_dat.drop(a1) print(seen_dat) sd = seen_dat.iloc[1:, 2] verify_model.fit(seen_dat, sd) inference = VariableElimination(verify_model) inference.induced_graph([ 'pen_pressure_f', 'letter_spacing_f', 'size_f', 'dimension_f', 'is_lowercase_f', 'is_continuous_f', 'slantness_f', 'tilt_f', 'entry_stroke_a_f', 'staff_of_a_f', 'formation_n_f', ' staff_of_d_f', 'exit_stroke_d_f ', 'word_formation', 'constancy' ]) phi_query = inference.map_query(variables=['pen_pressure_g'], evidence={ 'pen_pressure_f': 1, 'letter_spacing_g': 1, 'size_f': 2, 'is_lowercase_f': 2, 'slantness_f': 0, 'tilt_g': 1, 'entry_stroke_a_f': 1, 'staff_of_a_f': 0, 'formation_n_f': 1 }) print(phi_query)