EUVIMEDEUVIMED entdecken
Crossref

Adaptive Trade Based Money Laundering Detection in European Banking Networks Using Explainable Graph Learning and Sanctions Data

Shiba Prasad Biswas

Stout in Business and Management · 30.12.2025 · 1 (1) · 132-148

DOI: 10.61424/a3tb7v40 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Stout in Business and Management, online erschienen am 30.12.2025. Verfasst von Shiba Prasad Biswas.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Trade-based money laundering relies on the lawful nature of trade, settlement across borders and a disintegrated banking system, in which transaction-monitoring systems do not have a good visibility about the risk at the network-level. In this work, we propose an adaptive, interpretable graph-learning model to learn the most relevant suspicious pathways in the European banking network for TBML. To analyze the data, public FinCEN Files transaction-map records were processed, first for European-internal bank flows, then as temporal weighted graphs, and then compared to the European Union consolidated sanctions list, by isolating anomalies using anomaly learning and providing block-ablation explanations. This network comprised 2126 suspicious activity reports (SAR), 878 banks, 1035 bank pairs and USD 21.05 billion of reported flows. Adaptive ranking in 2016 gave a post-hoc sanctions-linkage ROC AUC 0.701 and top-five-percent review gave 3.47-fold enrichment compared to a frozen representation of 0.665. Explanations in the form of structural embeddings were predominant. The framework enhances the difference between suspicious reporting, sanctions risk and actual laundering, while giving better auditable prioritization of risks. Designed to be an investigative triage, not a legal adjudication.

Zitieren

APA

Biswas, S. P. (2025). Adaptive Trade Based Money Laundering Detection in European Banking Networks Using Explainable Graph Learning and Sanctions Data. Stout in Business and Management, 1(1), 132-148. https://doi.org/10.61424/a3tb7v40

Vancouver

Biswas SP. Adaptive Trade Based Money Laundering Detection in European Banking Networks Using Explainable Graph Learning and Sanctions Data. Stout in Business and Management. 2025;1(1):132-148. doi:10.61424/a3tb7v40

BibTeX
@article{Biswas2025,
  author = {Biswas, Shiba Prasad},
  title = {Adaptive Trade Based Money Laundering Detection in European Banking Networks Using Explainable Graph Learning and Sanctions Data},
  journal = {Stout in Business and Management},
  year = {2025},
  volume = {1},
  number = {1},
  pages = {132--148},
  doi = {10.61424/a3tb7v40}
}
Kostenlos starten →