return h2o_auc
if __name__ == '__main__':
parser = argparse.ArgumentParser(description='Modeling h2o')
parser.add_argument('--dataset',
'-d',
help="pass dataset name",
required=True)
parser.add_argument('--seed', '-s', help='random seed', default=2020)
args = parser.parse_args()
data_name = str(args.dataset)
seed = int(args.seed)
h2o.init()
X_train, X_test, y_train, y_test = load_data(data_name)
start_time = time.time()
score = h2o_train(X_train, X_test, y_train, y_test, seed=2020)
end_time = time.time()
hour = (end_time - start_time) / 3600.0
with open('h2o_result.txt', 'a') as f:
f.write(f"{data_name}\t{score}\t{hour}\n")
batch_size=batch_size),
dim=1)
logits = check_numpy(logits)
return logits
if __name__ == '__main__':
device = 'cuda' if torch.cuda.is_available() else 'cpu'
hours = 2 # 2 hours for training
res = []
for data_name, d in data_config.items():
# split dataset train/test = 0.7:0.3
X_train, X_test, y_train, y_test = load_data(data_name,
combine_y=False,
split_seed=2020,
test_size=0.3)
# general feature generator;
feature_generator = AutoMLFeatureGenerator()
print("#" * 50, 'training set preprocessing')
X_train = feature_generator.fit_transform(X_train,
drop_duplicates=False)
print("#" * 50, 'testing set preprocessing')
X_test = feature_generator.transform(X_test)
feature_types_metadata = feature_generator.feature_types_metadata
problem_type = 'binary'
path = f'LR-{data_name}'
"""
wide and deep test, follow code from autogluon
autogluon's NN architecture is based on wide and deep network
"""
from autogluon import TabularPrediction as task
from data_config.data_config import load_data, data_config
if __name__ == '__main__':
res = {}
for data_name in data_config.keys():
ylabel = data_config[data_name]['ylabel']
X_train, X_valid = load_data(data_name, combine_y=True)
train_data = task.Dataset(df=X_train)
test_data = task.Dataset(df=X_valid)
savedir = f'{data_name}/' # where to save trained models
predictor = task.fit(
train_data=train_data,
label=ylabel,
output_directory=savedir,
eval_metric='roc_auc',
verbosity=2,
visualizer='tensorboard',
random_seed=0,
save_space=True,
keep_only_best=True,
)
auc = predictor.evaluate(X_valid)
res[data_name] = auc
print(res)
import pickle
import torch
from sklearn.model_selection import StratifiedKFold
from data_config.data_config import data_config, load_data
import numpy as np
from sklearn.metrics import roc_auc_score
if __name__ == '__main__':
res = {}
for data_name in data_config.keys():
epoch = 100
train, test, y_train, y_test = load_data(data_name, combine_y=False)
types = train.dtypes
#
categorical_columns = []
categorical_dims = {}
features = list(train.columns)
print(train.shape)
for col in train.columns:
if types[col] == 'object':
train[col] = train[col].fillna("VV_likely")
test[col] = test[col].fillna('VV_likely')
d = train[col].unique()