Model-based fraud detection

“Data is everywhere, and with it are the clues to anticipate, prevent, and solve crimes; enhance security; and discover, detect, and deter unlawful and dangerous entities.” Jesus Mena (Investigative Data Mining for Security and Criminal Detection, 2003) Every day hundreds of thousand of people interact on technological platforms. These platforms offer cost-effective solutions for innovative digital services and customized user experience. as the cost of technology drops, it is envisioned that even more users will engage in online exchanges of information (transactions); and along with more transactions, attempts to exploit vulnerable technological platforms will have a huge impact on service-providing businesses and government agencies, not only in terms of direct monetary losses but also on the perceptions of active and potential users. Even a small percentage of fraudsters (i.e., users trying deceiving other parties) can quickly damage the reputation of a platform. This research proposes the development of a theoretical framework that exploits the stability properties of transactional patterns between transaction data to monitor their legitimacy.