Online auction frauds not as clever as they think, researchers say

        Criminals must think online auction sites are easy pickins based on the fact that fraud related to such sites accounted for about two-thirds of the 97,000 complaints referred to law enforcement agencies last year by the Internet Crime Complaint Center.

          Carnegie Mellon University computer scientists have developed software called Network Detection via Propagation of Beliefs (NetProbe) designed to make it harder for people to gain illegally from a multibillion online auction business that keeps getting bigger. 

         The researchers' software relies on data mining techniques for poring through publicly available records of transactions posted by sites such as eBay and identify suspicious online behaviors.   The data mining shows patterns where suspected frauds are seen to conduct business never or rarely with each other, for example.

         "To the best of our knowledge, this is the first work that uses a systematic approach to analyze and detect electronic auction frauds,” said Computer Science Professor Christos Faloutsos http://www.cs.cmu.edu/~christos/, in an official statement.  

           One test of about a million transactions by roughly 66,000 eBay users turned up 10 previously known fraudsters, plus a handful of probable ones and possible accomplices.           

        The researchers have used the software to come up with trustworthiness scores for individual user IDs, though have not made this publicly available to law enforcement or auction sites. This would complement user reputation scores, which can be manipulated by wrongdoers, as fraudsters team with others who conduct plenty of legitimate transactions to mask their involvement.

         Transactions on eBay are becoming fertile ground for network researchers. A University of Indiana study released earlier this year showed that phishers might be getting takers on as much as 14% of their trick messages, much higher than previous estimates by network security watchers. The university's School of Informatics simulated phishing attacks on eBay customers since they are such a popular target of actual online scams.

• Security