An Investigative Approach to Payment Card Fraud Detection using Machine Learning Techniques

Alu, Nwabuogoh Anne (2023) An Investigative Approach to Payment Card Fraud Detection using Machine Learning Techniques. Masters thesis, Dublin, National College of Ireland.

Preview	PDF (Master of Science) Download (1MB) | Preview
Preview	PDF (Configuration manual) Download (3MB) | Preview

Abstract

In today's rapidly evolving digital landscape, the threat of payment card fraud has escalated, imposing substantial financial burdens on individuals and businesses. This study is motivated by the imperative to counteract this menace, aiming to investigate the effectiveness of diverse supervised and deep learning techniques, including Extreme Gradient Boosting (XGBoost), Logistic Regression, LightGBM, Long Short-Term Memory Recurrent neural network (LSTMRNN), Random Forest and Multilayer Perceptron. To handle the problem of class imbalance, the hybrid approach was employed, SMOTE for oversampling and Edited Nearest Neighbors (ENN) for under-sampling. Notably, the model evaluation results highlight the prowess of boosting classifiers, especially LightGBM and XGBoost, in detecting credit card fraud, both techniques had an F1-Score of 0.63 and a PR-AUC score of 80% and 81.6% respectively.

Item Type:	Thesis (Masters)
Supervisors:	Name Email Byrne, Brian UNSPECIFIED
Uncontrolled Keywords:	LSTM; deep learning; LightGBM; MLP; Payment Card Fraud; Random Forrest; ANN; SMOTE
Subjects:	H Social Sciences > HG Finance > Banking H Social Sciences > HG Finance > Financial Services H Social Sciences > HG Finance > Fintech T Technology > T Technology (General) > Information Technology > Fintech Q Science > Q Science (General) > Self-organizing systems. Conscious automata > Machine learning
Divisions:	School of Computing > Master of Science in FinTech
Depositing User:	Tamara Malone
Date Deposited:	02 Aug 2024 10:16
Last Modified:	02 Aug 2024 10:16
URI:	https://norma.ncirl.ie/id/eprint/7012

Actions (login required)

View Item