Shill Bidding Detection in Online Auction

Abstract

AbstractThe acceptance of online auction systems has been increasing gradually over the years and this has led to a surge in many deceitful activities performed during transactions and functions in an online auction system (e.g. shill bidding, bid shielding, etc.). Exploitation of vulnerabilities in such systems for personal benefit in the past two decades has made winning the trust of customers a very difficult task. In this paper, we focus on the problem of shill bidding. To counteract shill bidding, we have used decision tree algorithm incorporating certain parameters which combine to point out who the shill bidder is in the online auction. At each stage we calculate the gain values, choosing the attribute with maximum value as the decisive factor for that particular pass. Each time a particular parameter for shill activity is detected it adds on to shill score and once this score exceeds our predefined boundary value that particular user is nabbed as a shill bidder. The admin can take appropriate steps to block this user in real-time to prevent him or her from participating in further auction process. Real time detection of the shill bidder as opposed to detection post auction proceedings is an aspect that has been looked into and worked upon.KeywordsOnline auctionDecision treeShill biddingE-commerceDetection