FinShield: Explainable Graph Neural Network Approach to Money Laundering Detection in Digital Social Transactions
P Ratna Tejaswi,
K Aanuj Reddy,
G Eesha,
J Nagalaxmi
The rapid growth of digital payment platforms and social network-based financial transactions has
increased the risk of money laundering activities. Criminals exploit peer-to-peer transfers, digital
wallets, and micro-transactions to disguise illicit funds. Traditional rule-based anti-money laundering
(AML) systems struggle to detect complex transaction patterns within social networks. This paper
proposes FinShield, an intelligent detection tool that leverages graph-based modeling and machine
learning techniques to identify suspicious transaction behaviors in social financial ecosystems. The
system constructs a user-transaction network and applies Graph Neural Networks (GNN) combined
with anomaly detection algorithms to detect laundering patterns. Experimental evaluation demonstrates
superior accuracy (95%) compared to traditional models, along with reduced false positive rates.
FinShield provides a scalable and adaptive framework for real-time AML monitoring.
