Augmenting Data with Generative Adversarial Networks to Improve Machine Learning-Based Fraud Detection

Autor(en): Fukas, Philipp
Menzel, Lukas
Thomas, Oliver 
Stichwörter: 'current; Adaptive boosting; Classification (of information); Crime; Data Augmentation; Financial Auditing; Financial fraud; Fraud Detection; Generative Adversarial Networks; Learning systems; Machine Learning; Machine learning methods; Machine learning models; Machine-learning; Prediction accuracy; Securities and Exchange Commissions; Support vector machines
Erscheinungsdatum: 2022
Herausgeber: Association for Information Systems
Journal: 17th International Conference on Wirtschaftsinformatik, WI 2022
Zusammenfassung: 
While current machine learning methods can detect financial fraud more effectively, they suffer from a common problem: dataset imbalance, i.e. there are substantially more non-fraud than fraud cases. In this paper, we propose the application of generative adversarial networks (GANs) to generate synthetic fraud cases on a dataset of public firms convicted by the United States Securities and Exchange Commission for accounting malpractice. This approach aims to increase the prediction accuracy of a downstream logit, support vector machine (SVM), and eXtreme Gradient Boosting (XGBoost) classifier by training on a more well-balanced dataset. While the results indicate that a state-of-the-art machine learning model like XGBoost can outperform previous fraud detection models on the same data, generating synthetic fraud cases before applying a machine learning model does not improve performance. © 2022 17th International Conference on Wirtschaftsinformatik, WI 2022. All rights reserved.
Beschreibung: 
Cited by: 0; Conference name: 17th International Conference on Wirtschaftsinformatik, WI 2022; Conference date: 21 February 2022 through 23 February 2022; Conference code: 191912
Externe URL: https://www.scopus.com/inward/record.uri?eid=2-s2.0-85172017387&partnerID=40&md5=e83574366249431a660ab089c3ff1267

Zur Langanzeige

Seitenaufrufe

1
Letzte Woche
0
Letzter Monat
0
geprüft am 14.05.2024

Google ScholarTM

Prüfen