Feature-Level Fusion of Super-App and Telecommunication Alternative Data\n Sources for Credit Card Fraud Detection

Identity theft is a major problem for credit lenders when there's not enough\ndata to corroborate a customer's identity. Among super-apps large digital\nplatforms that encompass many different services this problem is even more\nrelevant; losing a client in one branch can often mean losing them in other\nservices. In this paper, we review the effectiveness of a feature-level fusion\nof super-app customer information, mobile phone line data, and traditional\ncredit risk variables for the early detection of identity theft credit card\nfraud. Through the proposed framework, we achieved better performance when\nusing a model whose input is a fusion of alternative data and traditional\ncredit bureau data, achieving a ROC AUC score of 0.81. We evaluate our approach\nover approximately 90,000 users from a credit lender's digital platform\ndatabase. The evaluation was performed using not only traditional ML metrics\nbut the financial costs as well.\n