import os
from evalml.preprocessing import load_data
[docs]def load_fraud(n_rows=None, verbose=True, return_pandas=False):
"""Load credit card fraud dataset.
The fraud dataset can be used for binary classification problems.
Arguments:
n_rows (int): Number of rows from the dataset to return
verbose (bool): Whether to print information about features and labels
Returns:
Union[(ww.DataTable, ww.DataColumn), (pd.Dataframe, pd.Series)]: X and y
"""
currdir_path = os.path.dirname(os.path.abspath(__file__))
data_folder_path = os.path.join(currdir_path, "data")
fraud_data_path = os.path.join(data_folder_path, "fraud_transactions.csv.gz")
X, y = load_data(path=fraud_data_path,
index="id",
target="fraud",
n_rows=n_rows,
verbose=verbose)
X = X.set_types({"provider": "Categorical", "region": "Categorical"})
if return_pandas:
return X.to_dataframe(), y.to_series()
return X, y