Combining ranking concept and social network analysis to detect collusive groups in online auctions

► A Web crawling agent is used to collect real auction cases. ► The Page-Rank algorithm is used to discover the critical accounts of the groups. ► We developed a ranking method for auction fraud evaluation. ► The ANFIS neural network with real auction cases was implemented in the experiments. ► We find that the proposed ranking method is effective in identifying potential collusive fraud groups.