def train(resource_num, params, data_node):
print(resource_num, params)
start_time = time.time()
resource_num = resource_num * 1.0 / 27
# Prepare data node.
data_node = data_node['data_node']
_data_node = tmp_bo._parse(data_node, params)
X_train, y_train = _data_node.data
config_dict = params.copy()
# Prepare training and initial params for classifier.
init_params, fit_params = {}, {}
if _data_node.enable_balance == 1:
init_params, fit_params = get_fit_params(y_train, params['estimator'])
for key, val in init_params.items():
config_dict[key] = val
classifier_id, clf = get_estimator(config_dict)
try:
test_size = 0.2
score = partial_validation(clf, balanced_accuracy_scorer, X_train, y_train, resource_num,
test_size=test_size,
random_state=1,
if_stratify=True,
onehot=None,
fit_params=fit_params)
except Exception as e:
print(e)
score = -np.inf
print(resource_num, params, -score, time.time() - start_time)
# Turn it intos a minimization problem.
return {'loss': 1 - score, 'early_stop': False, 'lc_info': []}
def __call__(self, config, **kwargs):
start_time = time.time()
if self.name is None:
raise ValueError('This evaluator has no name/type!')
assert self.name in ['hpo', 'fe']
# Prepare configuration.
np.random.seed(self.seed)
config = config if config is not None else self.hpo_config
downsample_ratio = kwargs.get('data_subsample_ratio', 1.0)
# Prepare data node.
if 'data_node' in kwargs:
data_node = kwargs['data_node']
else:
data_node = self.data_node
X_train, y_train = data_node.data
config_dict = config.get_dictionary().copy()
# Prepare training and initial params for classifier.
init_params, fit_params = {}, {}
if data_node.enable_balance == 1:
init_params, fit_params = self.get_fit_params(
y_train, config['estimator'])
for key, val in init_params.items():
config_dict[key] = val
if data_node.data_balance == 1:
fit_params['data_balance'] = True
classifier_id, clf = get_estimator(config_dict)
if self.onehot_encoder is None:
self.onehot_encoder = OneHotEncoder(categories='auto')
y = np.reshape(y_train, (len(y_train), 1))
self.onehot_encoder.fit(y)
try:
if self.resampling_strategy == 'cv':
if self.resampling_params is None or 'folds' not in self.resampling_params:
folds = 5
else:
folds = self.resampling_params['folds']
score = cross_validation(
clf,
self.scorer,
X_train,
y_train,
n_fold=folds,
random_state=self.seed,
if_stratify=True,
onehot=self.onehot_encoder if isinstance(
self.scorer, _ThresholdScorer) else None,
fit_params=fit_params)
elif self.resampling_strategy == 'holdout':
if self.resampling_params is None or 'test_size' not in self.resampling_params:
test_size = 0.33
else:
test_size = self.resampling_params['test_size']
score = holdout_validation(
clf,
self.scorer,
X_train,
y_train,
test_size=test_size,
random_state=self.seed,
if_stratify=True,
onehot=self.onehot_encoder if isinstance(
self.scorer, _ThresholdScorer) else None,
fit_params=fit_params)
elif self.resampling_strategy == 'partial':
if self.resampling_params is None or 'test_size' not in self.resampling_params:
test_size = 0.33
else:
test_size = self.resampling_params['test_size']
score = partial_validation(
clf,
self.scorer,
X_train,
y_train,
downsample_ratio,
test_size=test_size,
random_state=self.seed,
if_stratify=True,
onehot=self.onehot_encoder if isinstance(
self.scorer, _ThresholdScorer) else None,
fit_params=fit_params)
else:
raise ValueError('Invalid resampling strategy: %s!' %
self.resampling_strategy)
except Exception as e:
if self.name == 'fe':
raise e
self.logger.info('%s-evaluator: %s' % (self.name, str(e)))
score = np.inf
self.logger.debug(
'%d-Evaluation<%s> | Score: %.4f | Time cost: %.2f seconds | Shape: %s'
% (self.eval_id, classifier_id, self.scorer._sign * score,
time.time() - start_time, X_train.shape))
self.eval_id += 1
# Turn it into a minimization problem.
return -score
def __call__(self, config, **kwargs):
start_time = time.time()
return_dict = dict()
self.seed = 1
downsample_ratio = kwargs.get('resource_ratio', 1.0)
# Prepare data node.
data_node = self.tmp_bo._parse(self.data_node, config)
X_train, y_train = data_node.data
config_dict = config.get_dictionary().copy()
# Prepare training and initial params for classifier.
init_params, fit_params = {}, {}
if data_node.enable_balance == 1:
init_params, fit_params = self.get_fit_params(
y_train, config['estimator'])
for key, val in init_params.items():
config_dict[key] = val
if data_node.data_balance == 1:
fit_params['data_balance'] = True
classifier_id, clf = get_estimator(config_dict)
if self.onehot_encoder is None:
self.onehot_encoder = OneHotEncoder(categories='auto')
y = np.reshape(y_train, (len(y_train), 1))
self.onehot_encoder.fit(y)
try:
if 'cv' in self.resampling_strategy:
if self.resampling_params is None or 'folds' not in self.resampling_params:
folds = 5
else:
folds = self.resampling_params['folds']
score = cross_validation(
clf,
self.scorer,
X_train,
y_train,
n_fold=folds,
random_state=self.seed,
if_stratify=True,
onehot=self.onehot_encoder if isinstance(
self.scorer, _ThresholdScorer) else None,
fit_params=fit_params)
elif 'holdout' in self.resampling_strategy:
if self.resampling_params is None or 'test_size' not in self.resampling_params:
test_size = 0.33
else:
test_size = self.resampling_params['test_size']
score = holdout_validation(
clf,
self.scorer,
X_train,
y_train,
test_size=test_size,
random_state=self.seed,
if_stratify=True,
onehot=self.onehot_encoder if isinstance(
self.scorer, _ThresholdScorer) else None,
fit_params=fit_params)
elif 'partial' in self.resampling_strategy:
if self.resampling_params is None or 'test_size' not in self.resampling_params:
test_size = 0.33
else:
test_size = self.resampling_params['test_size']
score = partial_validation(
clf,
self.scorer,
X_train,
y_train,
downsample_ratio,
test_size=test_size,
random_state=self.seed,
if_stratify=True,
onehot=self.onehot_encoder if isinstance(
self.scorer, _ThresholdScorer) else None,
fit_params=fit_params)
else:
raise ValueError('Invalid resampling strategy: %s!' %
self.resampling_strategy)
except Exception as e:
self.logger.info('evaluator: %s' % (str(e)))
score = -np.inf
# print(config)
# self.logger.info('%d-Evaluation<%s> | Score: %.4f | Time cost: %.2f seconds | Shape: %s' %
# (self.eval_id, classifier_id,
# self.scorer._sign * score,
# time.time() - start_time, X_train.shape))
self.eval_id += 1
# Turn it into a minimization problem.
return_dict['score'] = -score
return -score
def __call__(self, config, **kwargs):
start_time = time.time()
return_dict = dict()
if self.name is None:
raise ValueError('This evaluator has no name/type!')
assert self.name in ['hpo', 'fe']
# Prepare configuration.
np.random.seed(self.seed)
config = config if config is not None else self.hpo_config
downsample_ratio = kwargs.get('resource_ratio', 1.0)
# Prepare data node.
if 'data_node' in kwargs:
data_node = kwargs['data_node']
else:
data_node = self.data_node
X_train, y_train = data_node.data
config_dict = config.get_dictionary().copy()
regressor_id, reg = get_estimator(config_dict)
try:
if self.resampling_strategy == 'cv':
if self.resampling_params is None or 'folds' not in self.resampling_params:
folds = 5
else:
folds = self.resampling_params['folds']
score = cross_validation(reg,
self.scorer,
X_train,
y_train,
n_fold=folds,
random_state=self.seed,
if_stratify=False)
elif self.resampling_strategy == 'holdout':
if self.resampling_params is None or 'test_size' not in self.resampling_params:
test_size = 0.33
else:
test_size = self.resampling_params['test_size']
score = holdout_validation(reg,
self.scorer,
X_train,
y_train,
test_size=test_size,
random_state=self.seed,
if_stratify=False)
elif self.resampling_strategy == 'partial':
if self.resampling_params is None or 'test_size' not in self.resampling_params:
test_size = 0.33
else:
test_size = self.resampling_params['test_size']
score = partial_validation(reg,
self.scorer,
X_train,
y_train,
downsample_ratio,
test_size=test_size,
random_state=self.seed,
if_stratify=False)
else:
raise ValueError('Invalid resampling strategy: %s!' %
self.resampling_strategy)
except Exception as e:
if self.name == 'fe':
raise e
self.logger.info('%s-evaluator: %s' % (self.name, str(e)))
return -np.inf
self.logger.debug(
'%d-Evaluation<%s> | Score: %.4f | Time cost: %.2f seconds | Shape: %s'
% (self.eval_id, regressor_id, self.scorer._sign * score,
time.time() - start_time, X_train.shape))
self.eval_id += 1
# Turn it into a minimization problem.
return_dict['score'] = -score
return -score