Framework for Adaptive Sequential Pattern Recognition Applied on Credit Card Fraud Detection in the Online Games Industry

Michael Schaidnagel, Thomas Connolly, Fritz Laux

Research output: Contribution to journalArticle

Abstract

Online credit card fraud presents a significant challenge in the field of eCommerce. In 2012 alone, the total loss due to credit card fraud in the US amounted to $ 54 billion. Especially online games merchants have difficulties applying standard fraud detection algorithms to achieve timely and accurate detection. This paper describes the special constrains of this domain and highlights the reasons why conventional algorithms are not quite effective to deal with this problem. Our suggested solution for the problem originates from the fields of feature construction joined with the field of temporal sequence data mining. We present feature construction techniques, which are able to create discriminative features based on a sequence of transaction and are able to incorporate the time into the classification process. In addition to that, a framework is presented that allows for an automated and adaptive change of features in case the underlying pattern is changing.
Original languageEnglish
Pages (from-to)422-434
Number of pages13
JournalInternational Journal On Advances in Software
Volume7
Issue number3 and 4
Publication statusPublished - 2014

Fingerprint

Pattern recognition
Data mining
Industry

Keywords

  • feature construction
  • temporal data mining
  • binary classification
  • credit card fraud

Cite this

@article{c0d318c6ec754f44b2e6ebd665e95d8f,
title = "Framework for Adaptive Sequential Pattern Recognition Applied on Credit Card Fraud Detection in the Online Games Industry",
abstract = "Online credit card fraud presents a significant challenge in the field of eCommerce. In 2012 alone, the total loss due to credit card fraud in the US amounted to $ 54 billion. Especially online games merchants have difficulties applying standard fraud detection algorithms to achieve timely and accurate detection. This paper describes the special constrains of this domain and highlights the reasons why conventional algorithms are not quite effective to deal with this problem. Our suggested solution for the problem originates from the fields of feature construction joined with the field of temporal sequence data mining. We present feature construction techniques, which are able to create discriminative features based on a sequence of transaction and are able to incorporate the time into the classification process. In addition to that, a framework is presented that allows for an automated and adaptive change of features in case the underlying pattern is changing.",
keywords = "feature construction, temporal data mining, binary classification, credit card fraud",
author = "Michael Schaidnagel and Thomas Connolly and Fritz Laux",
year = "2014",
language = "English",
volume = "7",
pages = "422--434",
journal = "International Journal On Advances in Software",
issn = "1942-2628",
publisher = "International Academy, Research, and Industry Association",
number = "3 and 4",

}

Framework for Adaptive Sequential Pattern Recognition Applied on Credit Card Fraud Detection in the Online Games Industry. / Schaidnagel, Michael; Connolly, Thomas; Laux, Fritz.

In: International Journal On Advances in Software, Vol. 7, No. 3 and 4, 2014, p. 422-434.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Framework for Adaptive Sequential Pattern Recognition Applied on Credit Card Fraud Detection in the Online Games Industry

AU - Schaidnagel, Michael

AU - Connolly, Thomas

AU - Laux, Fritz

PY - 2014

Y1 - 2014

N2 - Online credit card fraud presents a significant challenge in the field of eCommerce. In 2012 alone, the total loss due to credit card fraud in the US amounted to $ 54 billion. Especially online games merchants have difficulties applying standard fraud detection algorithms to achieve timely and accurate detection. This paper describes the special constrains of this domain and highlights the reasons why conventional algorithms are not quite effective to deal with this problem. Our suggested solution for the problem originates from the fields of feature construction joined with the field of temporal sequence data mining. We present feature construction techniques, which are able to create discriminative features based on a sequence of transaction and are able to incorporate the time into the classification process. In addition to that, a framework is presented that allows for an automated and adaptive change of features in case the underlying pattern is changing.

AB - Online credit card fraud presents a significant challenge in the field of eCommerce. In 2012 alone, the total loss due to credit card fraud in the US amounted to $ 54 billion. Especially online games merchants have difficulties applying standard fraud detection algorithms to achieve timely and accurate detection. This paper describes the special constrains of this domain and highlights the reasons why conventional algorithms are not quite effective to deal with this problem. Our suggested solution for the problem originates from the fields of feature construction joined with the field of temporal sequence data mining. We present feature construction techniques, which are able to create discriminative features based on a sequence of transaction and are able to incorporate the time into the classification process. In addition to that, a framework is presented that allows for an automated and adaptive change of features in case the underlying pattern is changing.

KW - feature construction

KW - temporal data mining

KW - binary classification

KW - credit card fraud

M3 - Article

VL - 7

SP - 422

EP - 434

JO - International Journal On Advances in Software

JF - International Journal On Advances in Software

SN - 1942-2628

IS - 3 and 4

ER -