A Comparative Analysis of Machine Learning Algorithms for Credit Card Fraud Detection
Anshul Raturi, Maahi Pal, H.K. Narang, Satvik Vats, Vikrant Sharma, Satya Prakash Yadav
Abstract
The introduction and availability of credit cards in the modern world have transformed the transaction or payment process from the typical cash payment to modernized cashless payment. Due to the rapid growth and development of e-commerce platforms and improved accessibility to the Internet, a notable increase in the use of credit cards for purchases has occurred. Regrettably, this surge has also been accompanied by an upswing in fraudulent activities, demanding heightened vigilance. The phrase “credit card fraud” describes a wide rangeof illicit actions in which individuals utilize credit card details to get money through deception. Effectively combating this multifaceted threat necessitates a deep understanding of the mechanisms involved in identifying and preventing credit card fraud. Since sophisticated techniques for preventing and detecting fraud have been developed over time, the effects of credit card fraud have decreased, and the sector has become more stable. While previous research in this field has largely depended on machine learning algorithms such as random forest and logistic regression for credit card fraud detection, this study introduces a groundbreaking approach. We present the use of convolutional Neural network (CNN) models, a method commonly used in image processing and image recognition, to evaluate heatmaps generated from credit card transaction data.