dataset: https://www.kaggle.com/dalpozz/creditcardfraud
Detecting credit card fraud of a highly imbalanced dataset using SMOTE (synthetic minority oversampling technique) and ensemble learning (random forest) as well as a neural network
As much as I would have liked to work with the original dataset, for the sake of confidentiality this dataset is composed mostly of features which are the result of a PCA transformation of the original dataset. Some other features are the amount spent and time between transactions.