Credit Card Fraud Detection Using DNN

Abstract

Frauds in credit card transactions are common today and most happening as most of us are using the credit card payment methods more frequently. The reason behind it is the advancement in technology and surge in number of online transactions and corresponding financial loss. Therefore, there is need for effective and higher accuracy methods to reduce the loss. Moreover, fraudsters find ways to steal the credit card information of the user by sending fraud or fake SMS and calls, also through malicious attack, identity theft attack and so on. This paper aims in using the Deep Neural Networks algorithm of Deep Learning in predicting the occurrence of the fraud transaction. Further, we conduct a variation of the accomplished training and testing in deep learning techniques using balanced and imbalanced datasets to differentiate between fraud and non-fraud transactions and to acquire enough accuracy effectively.