In this repo I have used a basic neural net to predict whether a transaction is fraud or not. The data is highly imbalanced so a lot of pre processing has been applied to get better results.
Other classification algorithms may well be used to tweak results. My main objective here was to get an understanding of how to deal with highly imbalanced data and different techniques which can be applied to normalize the data and better the results.
Appreciate your time for viewing the repo.