Dataset for evaluate A2GAN paper
The dataset is pre-processed from two datasets: UCI credit card dataset (https://archive.ics.uci.edu/ml/datasets/default+of+credit+card+clients) and ULB credit card dataset (https://www.kaggle.com/mlg-ulb/creditcardfraud).
Using provided python scripts, you can generate several files with the format as follow: <A>_creditcard-<B>-<x.x>-<y.y>.csv, where
+ A is the name of dataset (uci or ulb)
+ B is train or test
+ x.x is percentage of missing data
+ y.y is percentage of missing label