Multi-channel data driven, real-time fraud determination system for electronic payment cards

• US 8,738,529 B2
• Filed: 07/15/2012
• Issued: 05/27/2014
• Est. Priority Date: 07/15/2011
• Status: Active Grant

• Current Assignee: Walmart Apollo, LLC (WalMart Inc.)
• Original Assignee: Wal Mart Stores Inc (WalMart Inc.)
• Inventors: Kolhatkar, Jayashree S., Fatnani, Sangita S., Yao, Yitao, Matsumoto, Kazuo
• Primary Examiner(s): Vezeris, James A
• Assistant Examiner(s): MADAMBA, CLIFFORD B
• Application Number:
  US13/549,491
• Publication Number:
  US 20130018795A1
• Time in Patent Office: 681 Days
• Field of Search: 705/44
• US Class Current: 705/44
• CPC Class Codes: G06Q 20/00 Payment architectures, schemes or protocols apparatus for performing or posting payment transactions G07F7/08, G07F19/00; electronic cash registers G07G1/12
• CPC Class Codes: G06Q 20/10 specially adapted for electronic funds transfer [EFT] systems; specially adapted for home banking systems
• CPC Class Codes: G06Q 20/4016 involving fraud or risk level assessment in transaction processing
• CPC Class Codes: G06Q 40/00 Finance; Insurance; Tax strategies; Processing of corporate or income taxes

Abstract

Exemplary embodiments for detecting electronic payment card fraud include receiving real-time payment card transaction data from ingress channels and an egress channels of at least one payment card system through a first application programming interface (API); generating transactional profiles for each of at least payment cards, the ingress channel, the egress channels, and funding sources of the payment cards; in response to receiving transaction data for a current payment card transaction, evaluating the transaction data using a predictive algorithm that compare the transaction data to the transactional profiles to calculate a probabilistic fraud score for the current transaction; evaluating the probabilistic fraud score and the current transaction data based on a set of rules to generate a recommendation to approve, decline or review the current transaction; and transmitting the recommendation back to the payment card system via a second API.

Claims

1. A computer-implemented method for detecting electronic payment card fraud, comprising:

receiving, by at least one software component executing on at least one processor, real-time payment card transaction data from ingress channels and an egress channels of at least one payment card system through a first application programming interface (API);

generating, by the at least one software component, at least two types of transactional profiles for each of at least payment cards, the ingress channel, the egress channels, and funding sources of the payment cards,

wherein a first type of transactional profile comprises a network graph of links between the ingress channels, the egress channels, the payment cards, and any users of the payment cards and geo-location identities associated with the payment cards; and

wherein a second type of transactional profile comprises a series of transactional usage patterns at different time intervals of each of the payment cards, the ingress channels, the egress channels, and the funding sources;

in response to the at least one software component receiving transaction data for a current payment card transaction, evaluating the transaction data using a predictive algorithm that compare the transaction data to the transactional profiles to calculate a probabilistic fraud score for the current transaction;

evaluating the probabilistic fraud score and the current transaction data based on a set of rules to generate a recommendation to approve, decline or review the current transaction; and

transmitting the recommendation back to the payment card system via a second API.

2. The method of claim 1 wherein generating transactional profiles further comprises populating an external data verification database with data retrieved from external sources and using the data for verification of geo-location, phone numbers, and addresses associated with the current payment card transaction.

3. The method of claim 1 further comprising calculating the probabilistic fraud score using multiple predictive algorithms comprising any combination of regression, decision trees, neural networks, random forest, and genetic algorithms.

4. The method of claim 1 wherein evaluating the probabilistic fraud score further comprises utilizing a rules language to build and deploy cutoff strategies for creating appropriate recommendations.

5. The method of claim 1 wherein transmitting the recommendation back to the payment card system further comprises one of embedding the recommendation as part of a response back to an incoming transaction request and posting the recommendation asynchronously to a pre-defined location.

6. The method of claim 1 wherein the method is performed by a fraud detection system that communicates with one or more card payments systems using the first and second APIs.

7. The method of claim 1 wherein the first API comprises a fixed component that allows for uniformity of sending data across multiple channels and a flexible component that allows each of the channels to send data that is unique to that channel.

8. An executable software product stored on a non-transitory computer-readable medium containing program instructions for detecting electronic payment card fraud, the program instructions for:

receiving real-time payment card transaction data from ingress channels and an egress channels of at least one payment card system through a first application programming interface (API);

generating at least two types of transactional profiles for each of at least payment cards, the ingress channel, the egress channels, and funding sources of the payment cards,

wherein a first type of transactional profile comprises a network graph of links between the ingress channels, the egress channels, the payment cards, and any users of the payment cards and geo-location identities associated with the payment cards; and

wherein a second type of transactional profile comprises a series of transactional usage patterns at different time intervals of each of the payment cards, the ingress channels, the egress channels, and the funding sources;

in response to receiving transaction data for a current payment card transaction, evaluating the transaction data using a predictive algorithm that compare the transaction data to the transactional profiles to calculate a probabilistic fraud score for the current transaction;

evaluating the probabilistic fraud score and the current transaction data based on a set of rules to generate a recommendation to approve, decline or review the current transaction; and

transmitting the recommendation back to the payment card system via a second API.

9. The software product of claim 8 wherein generating transactional profiles further comprises software instructions for populating an external data verification database with data retrieved from external sources and using the data for verification of geo-location, phone numbers, and addresses associated with the current payment card transaction.

10. The software product of claim 8 further comprising software instructions for calculating the probabilistic fraud score using multiple predictive algorithms comprising any combination of regression, decision trees, neural networks, random forest, and genetic algorithms.

11. The software product of claim 8 wherein evaluating the probabilistic fraud score further comprises software instructions for utilizing a rules language to build and deploy cutoff strategies for creating appropriate recommendations.

12. The software product of claim 8 wherein transmitting the recommendation back to the payment card system further comprises software instructions for one of embedding the recommendation as part of a response back to an incoming transaction request and posting the recommendation asynchronously to a pre-defined location.

13. The software product of claim 8 wherein the software instructions comprise a fraud detection system that communicates with one or more card payments systems using the first and second APIs.

14. The software product of claim 8 wherein the first API comprises a fixed component that allows for uniformity of sending data across multiple channels and a flexible component that allows each of the channels to send data that is unique to that channel.

15. A system, comprising:

a memory;

at least one processor coupled to the memory;

a first application programming interface (API) executed by the processor that is configured to receive real-time payment card transaction data from ingress channels and an egress channels of at least one payment card system;

a fraud monitoring component executed by the processor that is configured to generate at least two types of transactional profiles for each of at least payment cards, the ingress channel, the egress channels, and funding sources of the payment cards,

wherein a first type of transactional profile comprises a network graph of links between the ingress channels, the egress channels, the payment cards, and any users of the payment cards and geo-location identities associated with the payment cards; and

wherein a second type of transactional profile comprises a series of transactional usage patterns at different time intervals of each of the payment cards, the ingress channels, the egress channels, and the funding sources;

a fraud scoring component executed by the processor that is configured to receive transaction data for a current payment card transaction, and evaluate the transaction data using a predictive algorithm that compare the transaction data to the transactional profiles to calculate a probabilistic fraud score for the current transaction;

a fraud recommendation component executed by the processor that is configured to evaluate the probabilistic fraud score and the current transaction data based on a set of rules to generate a recommendation to approve, decline or review the current transaction; and

a second API executed by the processor that is configured to transmit the recommendation back to the payment card system.

16. The system of claim 15 wherein the fraud monitoring component further populates an external data verification database with data retrieved from external sources and using the data for verification of geo-location, phone numbers, and addresses associated with the current payment card transaction.

17. The system of claim 15 wherein the fraud scoring component calculates the probabilistic fraud score using multiple predictive algorithms comprising any combination of regression, decision trees, neural networks, random forest, and genetic algorithms.

18. The system of claim 15 wherein the fraud scoring component evaluates the probabilistic fraud score utilizing a rules language to build and deploy cutoff strategies for creating appropriate recommendations.

19. The system of claim 15 wherein the recommendation is transmitted back to the payment card system by one of embedding the recommendation as part of a response back to an incoming transaction request and posting the recommendation asynchronously to a pre-defined location.

20. The system of claim 15 wherein the system comprises a fraud detection system that communicates with one or more card payments systems using the first and second APIs.

21. The system of claim 15 wherein the first API comprises a fixed component that allows for uniformity of sending data across multiple channels and a flexible component that allows each of the channels to send data that is unique to that channel.