Time series models improve financial crime detection in transaction networks
Leveraging Time-Series Foundation Models for Financial Crime Detection
Machine Learning
Summary
Detecting illegal financial activity is hard because suspicious transactions are rare and happen over time in complex networks. The authors developed a way to summarize each account’s transaction history using advanced time-series models and then use that summary to better understand individual transactions. Adding this time-based information helps spot crime more accurately and faster than previous methods. They tested this approach on several datasets and found it improves detection, especially when combined with network analysis techniques.
What this means in practice
- •For financial compliance teams: Improve anti-money laundering systems by integrating summarized temporal transaction data to detect suspicious activity more accurately and quickly.
- •For blockchain security teams: Enhance phishing detection on Ethereum by using time-series account behavior summaries combined with graph models for better node classification.
Authors
Danae Natalie Savvidi, H. Çağrı Bilgi, Kubilay Atasu, Arie van Deursen
Abstract
Financial crime detection requires identifying illicit activity within large transaction networks, where suspicious behaviour is rare and unfolds over time. Existing graph-based methods capture relationships among accounts but typically underuse the temporal behaviour encoded in account transaction histories. We introduce Temporal-Context Edge Enrichment (TCEE), a framework that addresses this gap by summarising each account's transaction history with a frozen time-series foundation model (TSFM), and enriching each transaction's representation with the relevant temporal context from these summaries. These enriched representations can be scored directly for edge classification, or fed into a graph neural network (GNN) as edge features to capture broader structural information. Across six AMLWorld datasets, TCEE improves PR-AUC over the strongest evaluated baseline on five by a relative 10.8% on average, while training 6.0x faster end-to-end. Feeding TCEE's representations into a GNN raises PR-AUC a further 4.8% on the four smaller datasets. On an Ethereum phishing dataset, we show that TSFM-generated account representations fed into a GNN are also effective for node classification, improving PR-AUC by 14.5% over the strongest baseline. As TSFMs continue to advance, our results suggest they can strengthen financial crime detection by capturing temporal signals that complement graph-based approaches.