• Version
• Download 23
• File Size 518.30 KB
• File Count 1
• Create Date 19 December 2024
• Last Updated 19 December 2024

FRAUD PULSE

P. Rajapandian 1, R. Ramyasri 2, A. Deepa Shivani 3

1Associate Professor, Department of Computer Applications, Sri Manakula Vinayagar Engineering College (Autonomous), Puducherry 605008, India,

rajapandian.mca@smvec.ac.in

2Post Graduate student, Department of Computer Applications, Sri Manakula Vinayagar Engineering College (Autonomous),

Puducherry 605008, India p.ramyasri2002@gmail.com

3Post Graduate student, Department of Computer Applications, Sri Manakula Vinayagar Engineering College (Autonomous),

Puducherry 605008, India deepashivaniaroumougam@gmail.com

Abstract
The proliferation of e-commerce and advanced communication technologies has made credit cards the dominant payment method globally, both online and offline. This surge in credit card usage, however, has concurrently fueled a significant rise in fraudulent activities, resulting in substantial financial losses for individuals and businesses alike. Fraudsters continuously adapt their tactics, posing a significant challenge for researchers developing effective fraud detection systems. The inherent class imbalance in credit card fraud datasets—where legitimate transactions vastly outnumber fraudulent ones— further complicates the development of accurate detection models. Given the crucial role credit cards play in modern economies, impacting households, businesses, and global enterprises, the need for robust fraud detection mechanisms is paramount. This research proposes a Gradient Boosting Classifier, a powerful machine learning technique, as a novel approach to address this critical issue. Our experimental results demonstrate the superior performance of the proposed method compared to other machine learning algorithms, achieving a training accuracy of 100% and a test accuracy of 91%. This high accuracy underscores the effectiveness of the Gradient Boosting Classifier in accurately identifying fraudulent credit card transactions, offering a significant contribution to mitigating the risks associated with this pervasive form of financial crime.

Keywords: Credit Card Fraud Detection, Gradient Boosting Classifier, Machine Learning ,Imbalanced Dataset, Data Preprocessing, Accuracy, Flask Web Application, Support Vector Machine, Fraud Pulse, Feature Selection