Abstract
Online auctions are increasingly used as a smart and efficient way to optimise the consumers’ and sellers’ utility. A recently active field of research is the detection of fraud in online auctions. One of the most difficult types of fraud to detect is collusive shill bidding, where multiple user accounts jointly drive up the bids in an auction. This paper revises the Collusive Shill Bidding Algorithm(CSBD) proposed by Majadi et al. (2019) to develop an algorithm that is applied to a data set from an online auction platform (TBAuctions). We find that our algorithm converges, that computation time can be significantly reduced by appropriate choice of parameters, and we identify Shill Bidding for this data set, although the accuracy of the algorithm cannot be tested because of lack of ground truth values for the data. The paper further discusses steps needed for application of the algorithm to (very) large data sets, using a multiple core server, which despite substantial reduction in computation time would still require too much time to foresee a rapid implementation in real-time.
Original language | English |
---|---|
Pages (from-to) | 1-20 |
Journal | Computational Economics |
Volume | 63 |
Issue number | 1 |
DOIs | |
Publication status | Published - 2024 |
Bibliographical note
Publisher Copyright:© 2022, The Author(s).
Keywords
- Auctions
- Belief propagation
- Markov random field
- Online auctions
- Shill bidding