Inference of Online Auction Shills Using Dempster-Shafer Theory

Abstract

We present a shilling behavior detection and verification approach for online auction systems. Assuming a model checking technique to detect shill suspects in real-time, we focus on how to verify shill suspects using Dempster-Shafer theory of evidence. To demonstrate the feasibility of our approach, we provide a case study using real eBay auction data. The analysis results show that our approach can detect shills and that using Dempster-Shafer theory to combine multiple sources of evidence of shilling behavior can reduce the number of false positive results that would be generated from a single source of evidence.