EUVIMEDEUVIMED entdecken
Crossref

Federated Learning Architectures for Privacy Preserving Financial Fraud Detection Systems

Favour . C. Ezeugboaja

Frontiers in Emerging Artificial Intelligence and Machine Learning · 22.12.2025 · 2 (12) · 72-85

DOI: 10.64917/feaiml/volume02issue12-07 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Frontiers in Emerging Artificial Intelligence and Machine Learning, online erschienen am 22.12.2025. Verfasst von Favour . C. Ezeugboaja. Die Arbeit wurde bisher 2-mal zitiert (Crossref).

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

The increasing complexity and intensity of cases of financial fraud, such as synthetic identity fraud and international money laundering, have become significant concerns for classic fraud detection solutions, especially under strict data privacy regulations such as GDPR or EU Artificial Intelligence Act guidelines. This study focuses on the very pressing need to pursue high fraud detection performance while simultaneously ensuring user data confidentiality for highly fragmented financial systems. The study uses federated learning (FL) designs to analyse interesting opportunities for possibly entirely decentralized machine learning functions among diverse financial agencies without any need to transfer raw user information among them at all. Using a multimodal research methodology consisting of systematic literature studies, design studies, simulation studies, stress studies, and regulatory investigations, this study comprehensively assesses FL’s effectiveness and privacy pertaining to 2025 regulatory norms. The empirical results prove that FL not only improves overall fraud detection capacities but also ensures decent secrecy on raw personal information, meeting modern regulatory norms while formulating tolerant computing and secrecy constraints. The result implies the utmost importance of continued monitoring and attention to security loopholes, governance structure definitions, and dedicated investments into privacy boosting technologies to tap into FL’s revolutionary positive change potency within finance domains. Therefore, this work provides recent information on FL's dissemination implementation into disjointed finance domains, meeting both theoretical knowledge and practicality pursuits.

Zitieren

APA

Ezeugboaja, F. C. (2025). Federated Learning Architectures for Privacy Preserving Financial Fraud Detection Systems. Frontiers in Emerging Artificial Intelligence and Machine Learning, 2(12), 72-85. https://doi.org/10.64917/feaiml/volume02issue12-07

Vancouver

Ezeugboaja FC. Federated Learning Architectures for Privacy Preserving Financial Fraud Detection Systems. Frontiers in Emerging Artificial Intelligence and Machine Learning. 2025;2(12):72-85. doi:10.64917/feaiml/volume02issue12-07

BibTeX
@article{Ezeugboaja2025,
  author = {Ezeugboaja, Favour . C.},
  title = {Federated Learning Architectures for Privacy Preserving Financial Fraud Detection Systems},
  journal = {Frontiers in Emerging Artificial Intelligence and Machine Learning},
  year = {2025},
  volume = {2},
  number = {12},
  pages = {72--85},
  doi = {10.64917/feaiml/volume02issue12-07}
}
Kostenlos starten →