Bayesian Quickest Detection of Credit Card Fraud

This paper addresses the risk of fraud in credit card transactions by developing a probabilistic model for quickest detection illegitimate purchases. Using optimal stopping theory, goal is to determine moment, known as disorder or time, at which continuously monitored process consumer’s exhibits due fraud, order return best trade-off between two sources cost: on one hand, time should be detected soon possible counteract illegal activities and minimize loss that banks, merchants consumers suffer; other frequency false alarms minimized avoid generating adverse effects cardholders limit operational costs issuers. The proposed approach allows us score consumers’ determine, rigorous, personalized manner, threshold with scores are compared establish whether purchase fraudulent.