CORTEXA
← Browse

Rahil Sharma

1 paper indexed

arxivcs.LGcs.AI2026-07-21

Toward Auditable Fraud Detection: Combining Graph Features, Model Explanations, and Agentic Case Investigation

Rahil Sharma

Fraud detection systems must scale with rising transaction volume while remaining explainable and reviewable. We study a layered pipeline on the PaySim dataset that combines a gradient-boosted classifier, graph-derived structural features, an autoencoder-based anomaly signal, Tre…

View free PDFSource page