Comparative Analysis of Deep Learning Techniques For Credit Card Fraud Detection

Today, many types of frauds are carried out in the credit and debit card transactions. These types of transactions are usually carried out due to the increasing popularity of online transactions. Therefore, there's need to detect the online frauds while using credit card. There is a danger of stealing credit card information for malpractices. This is sometimes done by making fake calls or sending fake SMS or by sending phishing mails. There are many algorithms that can detect frauds, such as machine learning. In this paper, we study the various aspects of machine learning to detect frauds. This paper deals with Local Outlier Factor, Isolation Forest and Convolution Neural Network techniques. We determined accuracy, precision, recall, F1 score and loss for these techniques. We achieved an accuracy of 99% for both deep learning and supervised machine learning techniques.