if val_avg > 0.4:
val = val_max
elif val_avg < 0.01:
val = val_min
else:
val = val_max
keys = {k: val for k in keys}
map_result.update(keys)
list_res = []
for key, new in map_score2.items():
new = map_result[key]
try:
label = map_dup[key]
except:
continue
pred = map_score2[key]
new_pred = new
list_res.append((label, pred, new_pred))
aaa = pandas.DataFrame(list_res, columns=['label', 'pred', 'new'])
from tfidf_k import calc_weight
from sklearn.metrics import log_loss, roc_auc_score
sw = calc_weight(aaa['label'].values)
print(ppp)
print(roc_auc_score(aaa['label'].values, aaa['pred'].values, sample_weight=sw))
print(log_loss(aaa['label'].values, aaa['pred'].values, sample_weight=sw))
print(roc_auc_score(aaa['label'].values, aaa['new'].values, sample_weight=sw))
print(log_loss(aaa['label'].values, aaa['new'].values, sample_weight=sw))
print('-------')
from logging import StreamHandler, DEBUG, Formatter, FileHandler
log_fmt = Formatter('%(asctime)s %(name)s %(lineno)d [%(levelname)s][%(funcName)s] %(message)s ')
from logging import getLogger
logger = getLogger(__name__)
handler = StreamHandler()
handler.setLevel('INFO')
handler.setFormatter(log_fmt)
logger.setLevel('INFO')
logger.addHandler(handler)
aaa = pandas.read_csv('clique_data.csv')
sample_weight = calc_weight(aaa['label'].values)
# , 'emax', 'emin'] # , # 'l_score', 'r_score', 'm_score'] #
use_cols = ['cnum', 'pred', 'new', 'vmax', 'vmin', 'vavg'] # , 'emax', 'emin']
use_cols = ['cnum', 'pred', 'vmax', 'vmin', 'vavg'] # , 'emax', 'emin']
#'l_num', 'r_num', 'm_num']
x_train = aaa[use_cols].values
y_train = aaa['label'].values
all_params = {'max_depth': [5], # [14],
'learning_rate': [0.02], # [0.06, 0.1, 0.2],
'n_estimators': [10000],
'min_child_weight': [1],
'colsample_bytree': [0.7],