How do credit card companies investigate fraud

USI project prevents credit card fraud

An algorithm developed at the University of Italian Switzerland (USI) is intended to detect and prevent fraudulent credit card transactions in the future. It was developed as part of a research project funded by the AXA Research Fund. According to the project supervisor and head of the Data Science Lab at USI, Prof. Antonietta Mira, the chances are good that both AXA and other credit card companies will work with the algorithm in the future.

"We have developed a system that works in two phases - a training phase and a subsequent classification phase," explains Dr. Bruno Buonaguidi from Prof. Mira’s team in a press release. In 2016, he won the corresponding post-doctoral call from the AXA Research Fund.

First, the algorithm studies both the behavior of credit card users and the characteristics of reported fraudulent transactions. He calculates an individual threshold for each person. It indicates that a transaction is likely to be either okay or fraudulent. A fraud alarm is triggered if, for example, the period between two transactions is very tight or if they are made with unusually large amounts and in different locations than usual.

According to the USI, the annual fraud damage to credit card companies worldwide amounts to around 25 billion dollars. It could double by 2025.

According to Antonietta Mira, “the prospects for application by our project partner and other credit card companies are good”. The algorithm has been tested on real data and compared with other common machine learning methods. Her team was able to show that the percentage of fraudulent transactions found exceeded that of false reports. mm