EUVIMEDEUVIMED entdecken
Crossref

Parallelized Louvain-Based Community Detection and AntiBenford Subgraph Mining for Financial Fraud Detection in Transaction Networks

Amit Purushottam Pimpalkar, Rachna K. Somkunwar, Soham P. Chinchalkar, Kartik M. Katakdound, Atharva S. Bhide, Yogeshwar M. Patil

Informatica · 16.12.2025 · 49 (4)

DOI: 10.31449/inf.v498i4.8298 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Informatica, online erschienen am 16.12.2025. Verfasst von Amit Purushottam Pimpalkar und 5 weiteren Autor:innen. Die Arbeit wurde bisher 2-mal zitiert (Crossref).

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Detection of financial fraud remains an issue of concern since there is always a dynamic nature in the illegal patterns of transactions that conceal themselves in massive banking systems. The paper is a hybrid system that integrates the use of the AntiBenford statistical deviation analysis and graph-based community detection to identify the most appropriate methods to identify suspicious behavior. A transaction network graph forms the basis of the approach. Monetary flows, or directed edges, exist between each entity, which are an account, bank or merchant. AntiBenford module identifies digit anomalies of transaction records, whereas the augmented Louvain community detection algorithm, which runs in parallel, identifies well-knit communities, which are indicative of money laundering or collusion. We evaluated the model using the IBM Transactions Anti Money Laundering (AML) dataset where we obtained an accuracy of 96.57 percent more than traditional machine learning, rule-based and statistical anomaly methods. The reliability and interpretability of the method are validated by ROC AUC and precision-recall analysis. Combining statistical anomaly assessing with graph mining, this paper provides a scalable, flexible remedy to network fraud.

Zitieren

APA

Pimpalkar, A. P., Somkunwar, R. K., Chinchalkar, S. P., Katakdound, K. M., Bhide, A. S., & Patil, Y. M. (2025). Parallelized Louvain-Based Community Detection and AntiBenford Subgraph Mining for Financial Fraud Detection in Transaction Networks. Informatica, 49(4). https://doi.org/10.31449/inf.v498i4.8298

Vancouver

Pimpalkar AP, Somkunwar RK, Chinchalkar SP, Katakdound KM, Bhide AS, Patil YM. Parallelized Louvain-Based Community Detection and AntiBenford Subgraph Mining for Financial Fraud Detection in Transaction Networks. Informatica. 2025;49(4). doi:10.31449/inf.v498i4.8298

BibTeX
@article{Pimpalkar2025,
  author = {Pimpalkar, Amit Purushottam and Somkunwar, Rachna K. and Chinchalkar, Soham P. and Katakdound, Kartik M. and Bhide, Atharva S. and Patil, Yogeshwar M.},
  title = {Parallelized Louvain-Based Community Detection and AntiBenford Subgraph Mining for Financial Fraud Detection in Transaction Networks},
  journal = {Informatica},
  year = {2025},
  volume = {49},
  number = {4},
  doi = {10.31449/inf.v498i4.8298}
}
Kostenlos starten →