Publicação
Addressing Class Imbalance in Credit Card Fraud Detection: A Study of Artificial Sample Size Injection with GAN
| Resumo: | Credit card fraud continues to be a significant global challenge, leading to substantial financial losses. While traditional machine learning methods have served as the foundation for fraud detection, recent advancements in deep learning offer promise in capturing intricate fraudulent patterns. However, the persistent challenge of class imbalance in fraud datasets undermines the effectiveness of these models. This paper addresses this challenge by exploring Generative Adversarial Networks (GANs) to generate synthetic data, aiming to mitigate class imbalance. Specifically, the study investigates the optimal sample size of synthetic instances injected into the classifier to improve detection performance. Through experimentation on benchmark credit card transaction datasets, the study aims to identify the most effective combination of real and generated fraud instances for robust detection. The document includes a comprehensive review of existing methodologies, outlines the proposed approach, presents experimental findings, and discusses implications for future research. By doing so, this research contributes to the ongoing efforts to combat credit card fraud effectively. |
|---|---|
| Autores principais: | Silva, Ana Marta Rodrigues Barbosa da |
| Assunto: | Credit card fraud Generative Adversarial Networks (GANs) Class imbalance Fraud detection SDG 8 - Decent work and economic growth SDG 16 - Peace, justice and strong institutions |
| Ano: | 2025 |
| País: | Portugal |
| Tipo de documento: | dissertação de mestrado |
| Tipo de acesso: | acesso aberto |
| Instituição associada: | Universidade Nova de Lisboa |
| Idioma: | inglês |
| Origem: | Repositório Institucional da UNL |
| Resumo: | Credit card fraud continues to be a significant global challenge, leading to substantial financial losses. While traditional machine learning methods have served as the foundation for fraud detection, recent advancements in deep learning offer promise in capturing intricate fraudulent patterns. However, the persistent challenge of class imbalance in fraud datasets undermines the effectiveness of these models. This paper addresses this challenge by exploring Generative Adversarial Networks (GANs) to generate synthetic data, aiming to mitigate class imbalance. Specifically, the study investigates the optimal sample size of synthetic instances injected into the classifier to improve detection performance. Through experimentation on benchmark credit card transaction datasets, the study aims to identify the most effective combination of real and generated fraud instances for robust detection. The document includes a comprehensive review of existing methodologies, outlines the proposed approach, presents experimental findings, and discusses implications for future research. By doing so, this research contributes to the ongoing efforts to combat credit card fraud effectively. |
|---|