Fuzzy rule optimization for online auction frauds detection based on genetic algorithm

  • PDF / 1,129,355 Bytes
  • 14 Pages / 439.37 x 666.142 pts Page_size
  • 55 Downloads / 226 Views

DOWNLOAD

REPORT


Fuzzy rule optimization for online auction frauds detection based on genetic algorithm Cheng-Hsien Yu · Shi-Jen Lin

Published online: 5 April 2013 © Springer Science+Business Media New York 2013

Abstract Due to the huge amount of users and low entrance cost of online auction, there are a lot of online fraud cases in online auction sites. According to the IC3 reports from 2003 to 2010, we can understand the fraud cases and victims are increasing rapidly year by year. To improve the prevention of online auction frauds, this research will propose a hybrid approach to detect the fraudster accounts to help the users to identify which seller is more dangerous. In this research, we use social network analysis to produce the behavior features and transform these features into fuzzy rules which can represent the detection rules. Then optimize the fuzzy rules by genetic algorithms to build the auction fraud detection model. For implementation, we collect the real auction data from the online auction site http://www.ruten.com.tw which is the most popular auction site in Taiwan. Finally, we use the proposed features and methodologies to detect the fraudster accounts and find out the detection models of them. We hope the result of this research can help the website administrators to detect the possible collusive fraud groups easier in online auction.

Keywords Online auction · Fraud detection · Fuzzy rule · Social network analysis · Genetic algorithm

C.-H. Yu · S.-J. Lin National Central University, Jhongli City, Taoyuan, Taiwan S.-J. Lin e-mail: [email protected] C.-H. Yu () China University of Technology, Taipei, Taiwan e-mail: [email protected]

170

C.-H. Yu, S.-J. Lin

1 Introduction Due to the huge amount of users and low entrance cost of online auction, there are a lot of online fraud cases in online auction sites. Auction related frauds (e.g. auction fraud and no-delivery fraud) are the top popular fraud in both U.S. and Taiwan. The IC3 (Internet Crime Complaint Center) report of 2010 shows that there are 303,809 Internet complaints and 121,710 referrals submitted. The non-delivery fraud holds 21.1 % and auction fraud holds 10.1 % of referred complains [13]. From 2003 to 2010, auction related frauds were continuing to stand the top position of Internet complaint comparing [8–13]. According to the 2010 report of the fraud prevention hotline “165” in Taiwan, the auction related frauds are also hold the top popularity [17]. To prevent the frauds, every online auction site builds the login identification mechanism to check out the fraud problems. Despite this, the fraudsters always can cheat the careless users because of the physical identification is hard to complement in current process. One of the most common fraud methods is to establish several auction accounts, and using these accounts to have transactions with each other for increase their reputations. Their purpose is using these accounts to earn money by cheating. The current reputation system is too simple and provides too less information to help users to pr