for att in range(max_attempts):
for index_grado_privacy in range(len(grado_privacy_list)):
grado_privacy = grado_privacy_list[index_grado_privacy]
df.compute_band_matrix(dim_finale=dim_finale,
nome_file_item=listaItem,
num_sensibile=num_sensibile,
plot=False)
dataframe_bandizzato = df.dataframe_bandizzato.copy()
for index_r in range(len(r_list)):
print(
"%.2f |File : %s | Index privacy : %s | Attempt : %s | Index R: %s"
% (time.time() - big_ben, index_nameFile,
index_grado_privacy, att, index_r))
r = r_list[index_r]
cahd = CAHDalgorithm.CAHDalgorithm(
df, grado_privacy=grado_privacy, alfa=alpha)
cahd.compute_hist()
hist_item = cahd.hist
if cahd.CAHD_algorithm(analysis=True, plot=False):
if cahd.lista_gruppi is not None:
QID = cahd.lista_gruppi[0].columns.tolist()
KL_Divergence = 0
for kl_attempt in range(kl_attempts):
QID_select = list()
while len(QID_select) < r and len(
QID_select) < len(QID):
temp = random.choice(QID)
if temp not in QID_select:
QID_select.append(temp)
all_value = KLDivergence.get_all_combination_of_n(
r)
while len(QID_select) < r:
temp = random.choice(QID)
if temp not in QID_select:
QID_select.append(temp)
QID_list_to_select.append(QID_select)
KLs = list()
time_list = list()
q_time_list = list()
exit_list = list()
for iii in range(q_value_attempts):
start_time = time.time()
q_value = np.random.uniform(min_q, max_q)
cahd = CAHDalgorithm.CAHDalgorithm(df,
grado_privacy=grado_privacy,
alfa=alpha,
q_value=q_value)
cahd.compute_hist()
hist_item = cahd.hist
if cahd.CAHD_algorithm(analysis=True, plot=False):
end_time_1 = time.time() - start_time
KL_Divergence = 0
for ii in range(kl_attempts):
all_value = KLDivergence.get_all_combination_of_n(r)
# get max value of sensibile data
item_sensibile = int(
max(hist_item.keys(),
key=(lambda k: hist_item[k])))
QID_select = QID_list_to_select[ii]
for valori in all_value:
actsc = KLDivergence.compute_act_s_in_c(