target = 'PositiveReview'
ratio = 0.8
reviews_data = h2oai.create_dataset_sync(dataPath)
# Split the data
reviews_split_data = h2oai.make_dataset_split(dataset_key=reviews_data.key,
output_name1=basename + "_train",
output_name2=basename + "_test",
target=target,
fold_col="",
time_col="",
ratio=ratio,
seed=1234)
train_key = h2oai.get_dataset_split_job(reviews_split_data).entity[0]
test_key = h2oai.get_dataset_split_job(reviews_split_data).entity[1]
# Reviews Default
dropped = [
'UserID', 'ProductId', 'Id', 'Summary', 'Score', 'HelpfulnessDenominator',
'HelpfulnessNumerator', 'ProfileName', 'Time'
]
knobs = [8, 2, 7]
reviews1 = h2oai.start_experiment_sync(experiment_name="Reviews NLP Big",
dataset_key=train_key,
testset_key=test_key,
target_col=target,
is_classification=True,
accuracy=knobs[0],
card_data = h2oai.create_dataset_sync(dataPath)
# Split the data
card_split_data = h2oai.make_dataset_split(
dataset_key = card_data.key
, output_name1 = basename + "_train"
, output_name2 = basename + "_test"
, target = target
, fold_col = ""
, time_col = ""
, ratio = ratio
, seed = 1234
)
train_key = h2oai.get_dataset_split_job(card_split_data).entity[0]
test_key = h2oai.get_dataset_split_job(card_split_data).entity[1]
# Card Default
knobs = [6, 4, 6]
card_default = h2oai.start_experiment_sync(
experiment_name = "Card Default"
, dataset_key = train_key
, testset_key = test_key
, target_col = target
, is_classification = True
, accuracy = knobs[0]
, time = knobs[1]
, interpretability = knobs[2]
, enable_gpus = True
, cols_to_drop = dropped
ratio = 0.8
diabetes_data = h2oai.create_dataset_sync(dataPath)
# Split the data
diabetes_split_data = h2oai.make_dataset_split(dataset_key=diabetes_data.key,
output_name1=basename +
"_train",
output_name2=basename + "_test",
target=target,
fold_col="",
time_col="",
ratio=ratio,
seed=1234)
train_key = h2oai.get_dataset_split_job(diabetes_split_data).entity[0]
test_key = h2oai.get_dataset_split_job(diabetes_split_data).entity[1]
dropped = []
# Diabetes Default
knobs = [8, 2, 8]
diabetes1 = h2oai.start_experiment_sync(experiment_name="Diabetes",
dataset_key=train_key,
testset_key=test_key,
target_col=target,
is_classification=True,
accuracy=knobs[0],
time=knobs[1],
interpretability=knobs[2],
enable_gpus=True,
cols_to_drop=dropped)
target = 'VALUE'
ratio = 0.8
boston_data = h2oai.create_dataset_sync(dataPath)
# Split the data
boston_split_data = h2oai.make_dataset_split(dataset_key=boston_data.key,
output_name1=basename + "_train",
output_name2=basename + "_test",
target=target,
fold_col="",
time_col="",
ratio=ratio,
seed=1234)
train_key = h2oai.get_dataset_split_job(boston_split_data).entity[0]
test_key = h2oai.get_dataset_split_job(boston_split_data).entity[1]
dropped = []
# Housing Experiment #1
knobs = [7, 2, 8]
housing1 = h2oai.start_experiment_sync(experiment_name="Housing",
dataset_key=train_key,
testset_key=test_key,
target_col=target,
is_classification=False,
accuracy=knobs[0],
time=knobs[1],
interpretability=knobs[2],
scorer='RMSE',
enable_gpus=True,
target = 'survived'
ratio = 0.8
titanic_data = h2oai.create_dataset_sync(dataPath)
# Split the data
titanic_split_data = h2oai.make_dataset_split(dataset_key=titanic_data.key,
output_name1=basename + "_train",
output_name2=basename + "_test",
target=target,
fold_col="",
time_col="",
ratio=ratio,
seed=1234)
train_key = h2oai.get_dataset_split_job(titanic_split_data).entity[0]
test_key = h2oai.get_dataset_split_job(titanic_split_data).entity[1]
knobs = [8, 2, 8]
# Titanic Default
dropped = ['no.title', 'cabin', 'embarked', 'boat', 'body', 'home.dest']
titanic1 = h2oai.start_experiment_sync(experiment_name="Titanic",
dataset_key=train_key,
testset_key=test_key,
target_col=target,
is_classification=True,
accuracy=knobs[0],
time=knobs[1],
interpretability=knobs[2],