Invariant Diversity as a Proactive Fraud Detection Mechanism for Online Merchants

Laurens, Roy, Jusak, Jusak ORCID: and Zou, Cliff C. (2017) Invariant Diversity as a Proactive Fraud Detection Mechanism for Online Merchants. In: IEEE Global Communications Conference (GLOBECOM), 4-8 December 2017, Singapore.

1. Dokumen Globecom2017.pdf - Accepted Version

Download--- (1MB) | Preview
2. Peer Review Blobecom17.pdf - Accepted Version

Download--- (546kB) | Preview
3. Turnitin GLOBECOM2017.pdf - Accepted Version

Download--- (2MB) | Preview

Search this title on : |


Online merchants face difficulties in using existing card fraud detection algorithms, so in this paper we propose a novel proactive fraud detection model using what we call invariant diversity to reveal patterns among attributes of the devices (computers or smartphones) that are used in conducting the transactions. The model generates a regression function from a diversity index of various attribute combinations, and use it to detect anomalies inherent in certain fraudulent transactions. This approach allows for proactive fraud detection using a relatively small number of unsupervised transactions and is resistant to fraudsters’ device obfuscation attempt. We tested our system successfully on real online merchant transactions and it managed to find several instances of previously undetected fraudulent transactions.

Export Record

Item Type: Conference or Workshop Item (Paper)
Additional Information: Roy Laurens, Jusak, Cliff C. Zou
Uncontrolled Keywords: Electronic Commerce; credit card fraud; fraud prevention; diversity index
Dewey Decimal Classification: 000 – Computer science, information & general works > 000 Computer science, knowledge & systems > 005 Computer programming, programs & data
Divisions: Perpustakaan > Prosiding/Call for Papers
Depositing User: Annuh Liwan Nahar
Date Deposited: 10 May 2021 08:48
Last Modified: 10 May 2021 08:48

Download Statistics

Downloads over the past year. Other digital versions may also be available to download e.g. from the publisher's website.

View more statistics

Actions (login required)

View Item   View Item