Abstract
This study develops methods to detect anomalous transactions linked with fraud in food stamp purchases through order statistics methods. The methods detect clusters in the order statistics of the transaction amounts that merit further scrutiny. Our techniques use scan statistics to determine when an excessive number of transactions occur (cluster), which is historically linked to fraud. A scoring paradigm is constructed that ranks the degree in which detected clusters and individual transactions are anomalous among approximately 250 million total transactions.
Funding Statement
Robert Lund was partially supported by NSF Grant DMS-2113592. This project was partially supported by the USDA grant SNAP-RIIT-2015 through Mississippi’s Department of Human Services.
Acknowledgments
The authors thank two anonymous referees and the Editors for constructive comments that improved this paper. We acknowledge work from Leigh Ellen Barefield and Sheida Riahi, and additionally acknowledge the resources and facilities at NSPARC.
Citation
Jonathan Woody. Zhicong Zhao. Robert Lund. Tung-Lung Wu. "A forensic statistical analysis of fraud in the federal food stamp program." Ann. Appl. Stat. 18 (3) 2486 - 2510, September 2024. https://doi.org/10.1214/24-AOAS1891
Information
