def calculate_metafeatures(task_id):
X_train, y_train, X_test, y_test, cat, task_type, dataset_name = load_task(task_id)
watch = StopWatch()
if task_type == 'classification':
if len(np.unique(y_train)) == 2:
task_type = BINARY_CLASSIFICATION
else:
task_type = MULTICLASS_CLASSIFICATION
else:
task_type = REGRESSION
_metafeatures_labels = _calculate_metafeatures(
x_train=X_train, y_train=y_train, data_feat_type=cat,
data_info_task=task_type, basename=dataset_name, logger_=logger,
watcher=watch,
)
_metafeatures_encoded_labels = _calculate_metafeatures_encoded(
x_train=X_train, y_train=y_train, data_feat_type=cat,
task=task_type, basename=dataset_name, logger_=logger,
watcher=watch,
)
mf = _metafeatures_labels
mf.metafeature_values.update(
_metafeatures_encoded_labels.metafeature_values)
return mf
def test_metalearning(self):
dataset_name_classification = 'digits'
initial_challengers_classification = {
"ACC_METRIC":
"--initial-challengers \" "
"-balancing:strategy 'weighting' "
"-classifier:__choice__ 'proj_logit'",
"AUC_METRIC":
"--initial-challengers \" "
"-balancing:strategy 'weighting' "
"-classifier:__choice__ 'liblinear_svc'",
"BAC_METRIC":
"--initial-challengers \" "
"-balancing:strategy 'weighting' "
"-classifier:__choice__ 'proj_logit'",
"F1_METRIC":
"--initial-challengers \" "
"-balancing:strategy 'weighting' "
"-classifier:__choice__ 'proj_logit'",
"PAC_METRIC":
"--initial-challengers \" "
"-balancing:strategy 'none' "
"-classifier:__choice__ 'random_forest'"
}
dataset_name_regression = 'diabetes'
initial_challengers_regression = {
"A_METRIC":
"--initial-challengers \" "
"-imputation:strategy 'mean' "
"-one_hot_encoding:minimum_fraction '0.01' "
"-one_hot_encoding:use_minimum_fraction 'True' "
"-preprocessor:__choice__ 'no_preprocessing' "
"-regressor:__choice__ 'random_forest'",
"R2_METRIC":
"--initial-challengers \" "
"-imputation:strategy 'mean' "
"-one_hot_encoding:minimum_fraction '0.01' "
"-one_hot_encoding:use_minimum_fraction 'True' "
"-preprocessor:__choice__ 'no_preprocessing' "
"-regressor:__choice__ 'random_forest'",
}
for dataset_name, task, initial_challengers in [
(dataset_name_regression, REGRESSION,
initial_challengers_regression),
(dataset_name_classification, MULTICLASS_CLASSIFICATION,
initial_challengers_classification)
]:
for metric in initial_challengers:
configuration_space = get_configuration_space(
{
'metric': metric,
'task': task,
'is_sparse': False
},
include_preprocessors=['no_preprocessing'])
X_train, Y_train, X_test, Y_test = get_dataset(dataset_name)
categorical = [False] * X_train.shape[1]
meta_features_label = _calculate_metafeatures(
X_train, Y_train, categorical, dataset_name, task)
meta_features_encoded_label = _calculate_metafeatures_encoded(
X_train, Y_train, categorical, dataset_name, task)
initial_configuration_strings_for_smac = \
suggest_via_metalearning(
meta_features_label,
meta_features_encoded_label,
configuration_space, dataset_name, metric,
task, False, 1, None)
print(metric)
print(initial_configuration_strings_for_smac[0])
self.assertTrue(
initial_configuration_strings_for_smac[0].startswith(
initial_challengers[metric]))