def plot3(alphabet_sizes,
data_size=400,
missingness_parameter=0.1,
dl="me",
minsup=0.02):
similarity_list = []
for alphabet_size in alphabet_sizes:
positives = []
negatives = []
while len(positives) == 0 or len(negatives) == 0:
actual_theory = generate_theory(alphabet_size)
theory_object = GeneratedTheory(actual_theory)
generator = TheoryNoisyGeneratorOnDataset(theory_object, data_size,
missingness_parameter)
positives, negatives = generator.generate_dataset()
learned_theory, _ = Mistle(positives, negatives).learn(dl_measure=dl,
minsup=minsup)
learned_clauses = get_clauses(learned_theory)
similarity = compare_theories(actual_theory, learned_clauses,
alphabet_size)
similarity_list.append(similarity)
# print(similarity_list)
return similarity_list
def plot1(data_sizes,
alphabet_size=14,
missingness_parameter=0.1,
dl="me",
minsup=0.02):
similarity_list = []
for data_size in data_sizes:
positives = []
negatives = []
while len(positives) == 0 or len(negatives) == 0:
actual_theory = []
while actual_theory == []:
actual_theory = generate_complex_theory(
alphabet_size, clause_size=(2, alphabet_size))
if solve([tuple(clause)
for clause in actual_theory]) == "UNSAT":
print("Generated theory is not SAT\t: " +
str(actual_theory))
actual_theory = []
theory_object = GeneratedTheory(actual_theory)
generator = TheoryNoisyGeneratorOnDataset(theory_object, data_size,
missingness_parameter)
positives, negatives = generator.generate_dataset()
learned_theory, _ = Mistle(positives, negatives).learn(dl_measure=dl,
minsup=minsup)
learned_clauses = get_clauses(learned_theory)
similarity = compare_theories(actual_theory, learned_clauses,
alphabet_size)
similarity_list.append(similarity)
return similarity_list
samp_positives_list = []
samp_negatives_list = []
samp_unique_positives_list = []
samp_unique_negatives_list = []
overlap_size_list = []
# simple_theory_list = []
# for i in range(iterations):
# simple_theory_list.append(generate_theory(alphabet_size))
for i in range(iterations):
print("\n\n==================== New Iteration [" + str(i) +
"] ====================")
actual_theory = complex_theory_list[i]
# actual_theory = simple_theory_list.pop()
theory_object = GeneratedTheory(actual_theory)
generator = TheoryNoisyGeneratorOnDataset(theory_object, data_size,
missingness_parameter)
positives, negatives = generator.generate_dataset()
samp_positives = len(positives)
samp_negatives = len(negatives)
samp_unique_positives = len(set(positives))
samp_unique_negatives = len(set(negatives))
if samp_positives == 0:
print(i, "Sampled Positives are empty")
if samp_negatives == 0:
print(i, "Sampled Negatives are empty")
if balance: