def get_list_tf_on_tran(self, csv_name):
rows = []
for file_name in self.test_files_list:
index_file = self.get_tran_index_fromFileName(file_name=file_name)
# for start,end in self.dataset_answer_st_ed_list[index_file]:
c = self.change_points[index_file]
temp = c
# temp = self.gen_tf_from_changepoint(start, end, self.change_points[index_file])
temp.append(file_name)
rows.append(temp)
ut.list_to_csv(rows=rows, csv_name=csv_name)
resultRow.append(false_alerts)
resultRow.append(total_alert)
resultRow.append(pos)
resultRow.append(neg)
tp_rate = float(true_alerts) / pos
resultRow.append(tp_rate)
fp_rate = float(false_alerts) / neg
resultRow.append(fp_rate)
resultRow.append(sum(found))
return resultRow
if __name__ == '__main__':
ut.list_to_csv(head=head)
result = []
df_list = []
name_list = ut_light.gen_list_mix(main_path=dataset_path,data_type='bgs',mix_size="*sq_L3_I5_*")
for index, name in enumerate(name_list):
for threshold in thresholds:
algo_name = "SEA_{}_{}".format(bin_sizes,block_type)
ut.checkFolderandCreate("{}\\ROC\\{}".format(main_path, algo_name))
pickle_ROC_file = '{}\\ROC\\{}\\_{}_threshold{}.pkl'.format(main_path, algo_name, name, threshold)
if not os.path.exists(pickle_ROC_file):
result = []
data = ut_light.load_light_select_dataset(input_path=dataset_path, file_name=name, is_mix=is_mix)
print("##### cal name : {} , threshold = {} #####".format(name, threshold))
ROC_row = genROCFile(name=name, data=data, threshold=threshold)
result.append(ROC_row)
df = pd.DataFrame(result, columns=ROC_head)
for pattern in patterns:
for I in IS:
# path = 'D:\\git_project\\data stream\\lightcurve_benchmark\\{}'.format(data_type)
path = 'C:\\git\\data_stream\\lightcurve_benchmark\\{}'.format(data_type)
# path = 'C:\\git\\data_stream\\gaia\\{}'.format(data_type)
data = data_stream(path=path, type=type,pattern=pattern, len=L,interval=I)
data.load_data_fromfile()
for bin_period in bin_periods:
file_format_name = '{}_{}_bin{}'.format(type, data_type, bin_period)
for k in ks:
data_sk, result_text = cal_sk_bin(data=data,
input_bin=bin_period,input_k=k,
file_format="{}_k{}".format(file_format_name,k), L=L, I=I)
print("{}_k{} cheb sk".format(file_format_name,k))
head, rows = data_sk.result_for_ROC(data_type, Algo='SK',K=k, Bin=bin_period, cheb_size=Cheb_win)
ut.list_to_csv(rows=rows, csv_name="ks_bin//roc_bin.csv", is_append=True)
head, rows = data_sk.result_for_ROC(data_type, Algo='SK', K=k,
Bin=bin_period,is_normalization=True,cheb_size=Cheb_win)
ut.list_to_csv(rows=rows, csv_name="ks_bin//roc_nor.csv", is_append=True)
#
# # ut.list_to_txt(rows=result_text, csv_name="ks_bin//{}".format(txt_file_name),
# # is_append=True)
data_bin_k,result_text = cal_CANDAR_bin(data=data,
input_bin=bin_period,input_k=k,
file_format="{}_k{}".format(file_format_name,k)
, L=L,I=I)
print("{}_k{} cheb k".format(file_format_name,k))
head, rows = data_bin_k.result_for_ROC(data_type, Algo='CHEB_K', K=k, Bin=bin_period, cheb_size=Cheb_win)
ut.list_to_csv(rows=rows, csv_name="CANDAR_bin//roc_bin.csv", is_append=True)
head, rows = data_bin_k.result_for_ROC(data_type, Algo='CHEB_K', K=k, Bin=bin_period,
is_normalization=True,cheb_size=Cheb_win)
types = ["Ligthcurve"]
patterns = ["sq"]
LS = [5]
IS = [30]
for data_type in data_types:
for k in ks:
for type in types:
for pattern in patterns:
for L in LS:
for I in IS:
data = data_stream(
# path='D:\\git_project\\data stream\\lightcurve_benchmark\\{}'.format(data_type), type=type,
# pattern=pattern, len=L,
# interval=I)
path='C:\\git\\data_stream\\lightcurve_benchmark\\{}'.format(data_type), type=type,
pattern=pattern, len=L,
interval=I)
data.load_data_fromfile()
data = cal_k_tran(data=data, input_k=k)
# # Cheb
print("### end change detection##")
head,rows,csv_path = data.result_for_ROC_old(filter_type=data_type, Algorithm='Kalman', K=k)
ut.list_to_csv(rows=rows,csv_name=csv_path,is_append=True)
# data.save_f1_to_csv( L=L, I=I,
# Algorithm="Base Cheb",
# Dataset_type=type, pattern=pattern,K=k,Bin="None")
