Systems and methods for early detection of network fraud events
Abstract:
A computing system for detecting fraudulent payment card network events includes a ratio striping engine that receives scored payment card transaction authorization requests and generates data structures for a plurality of account ranges. Each data structure sorts the transaction authorization requests within the associated account range over a plurality of fraud score range stripes. The data structures are parsed over time periods that extend back from a common starting point. For each data structure and time period, at least one cumulative metric is calculated from the transaction authorization requests in each fraud score range stripe. For each data structure, ratio striping values are determined between values of the at least one metric in a fraud score range stripe over two of the time periods. A fraud event associated with at least one of the account ranges is detected based on the ratio striping values for the corresponding data structure.
Public/Granted literature
Information query
Patent Agency Ranking
0/0