The objective of this project is to apply autoencoders to detect frauds in a credit card dataset. The dataset used for the project can be found here: https://www.kaggle.com/mlg-ulb/creditcardfraud
We will use the dataset as follows:
- 199,659 non-fraudulent transactions (Class 0) to train the autoencoders
- 50,000 records with 91 fraudulent transactions as validation set to define the error threshold
- 34,807 records with 60 fraudulent transactions as test set to make predictions