Effective Appeal of Fraud Detection in Credit Card Data Using Machine Learning Techniques

Effective Appeal of Fraud Detection in Credit Card Data Using Machine Learning Techniques

K. Geetha Rani¹, Chippe Ashok², Haalini Mandava³, R. Sai Pranav⁴, Santati Uttej⁵

²³⁴⁵Final Year B.Tech Student, Department of CSE, Jain University, Bengaluru, Karnataka, India.

¹Associate Professor,  Department of CSE, Jain University, Bengaluru, Karnataka, India.

Abstract
Banks and other financial organizations are extremely concerned about credit card theft on a global scale.More and more, real-time fraud detection is being done using machine learning algorithms. In this study, two wellknown machine learning techniques—Light Gradient Boosting Machine and Isolation Forest—are used to investigate the detection of credit card fraud. These algorithms were utilised in the study to analyse transactional data and look for patterns of credit card fraud. The results showed that both algorithms had high rates of success in identifying fraudulent behaviour. The Isolation Forest approach outperformed the Light Gradient Boosting Machine technique in problems involving classification, regression, and ranking. The authors conclude that utilising machine learning techniques can significantly improve by using machine learning techniques.

Key Words: Light Gradient Boosting Method, Isolation Forest.