User action-based financial fraud detection method by SVDD

Abstract

This thesis proposes a method to detect sophisticated electronic financial frauds using SVDD. The financial industry detects electronic financial frauds using FDS, but its false positive rate is high enough to require additional authentications. It causes customers inconveniences and does not detect those sophisticated financial frauds. In order to resolve the aforementioned issues, this study proposes a method to detect such potential frauds by profiling and vectorizing user activities and device information by SVDD.