Financial Fraud Alert Review Dataset Description

This research article introduces the Financial Fraud Alert Review Dataset (FiFAR), a synthetic dataset designed to enhance learning to defer (L2D) algorithms in financial fraud detection. It addresses the limitations of existing public datasets by providing predictions from 50 synthetic fraud analysts, enabling robust benchmarking of human-AI collaboration in decision-making systems. The dataset is aimed at fostering research in L2D methods under realistic conditions, promoting effective human-A