Machine-Learning-Based Scoring System for Antifraud CISIRTs in Banking Environment

The number of fraud occurrences in electronic banking is rising each year. Experts the field cybercrime are continuously monitoring and verifying network infrastructure transaction systems. Dedicated threat response teams (CSIRTs) used by organizations to ensure security stop cyber attacks. Financial institutions well aware this have increased funding for CSIRTs antifraud software. If company has a rule-based system, CSIRT can examine cases create rules counter threat. not, they attempt analyze Internet traffic down packet level look anomalies before adding proxy or firewall servers mitigate However, does not always solve issues, because transactions occasionally receive “gray” rating. Nevertheless, bank unable approve every gray call center employees insufficient make possible. In study, we designed machine-learning-based rating system that provides early warnings against financial fraud. We present architecture together with new ML-based scoring extension, which examines customer logins from system. suggested method enhances organization’s prevention Because occur immediately after client identification authorization process, quickly identify operations. reduces amount successful improves queue administration.