AuctionTalk Radio logo

 

Contact info:

Interested in being
on the show?

Email us

or call us at:
818-317-0354

 

This site is best viewed in Mozilla Firefox or Safari. Internet Explorer CSS layout problems are not supported.

Auction Talk Radio - what's the story of your stuff?

Welcome to Auction Talk Radio

Auction Talk Radio is intended to be America's first radio show dedicated to auctions, collectibles and eBay. If you are an eBay enthusiast, a garage sale junkie, an antique lover, or just simply want to know what your old stuff is worth, you won't want to miss this show. We were recently guests on KNX Radio here in L.A. and may soon hit the airwaves where you live. Welcome to our website!

 

New Tool Unveiled in Fight vs. Net Auction Fraud


Carnegie Mellon NetProbe data-mines for "false positives"

Researchers at Carnegie Mellon University have developed software that may be the next step forward in the detection of online auction crooks. The project, called Network Detection via Propagation of Beliefs, or NetProbe, is designed to combat those who puff up their reputation ratings (eBay's feedback score, for instance) in an effort to represent themselves as trustworthy buyers and sellers. The CMU program looks for patterns of users who have repeated dealings with one another, and alerts other users that there is a higher probability of having a fraudulent transaction with them. The university announced the project last month as a news item called "Carnegie Mellon Researchers Uncover Online Auction Fraud" on its web site, with myriad media picking up the story this week, including Yahoo! Finance.

Revealing what many experienced online auctioneers already know too well, the CMU researchers said savvy fraudsters conduct transactions with "friends or even themselves, using alternate user names to give themselves high satisfaction ratings - so unsuspecting customers will still try to buy from them." The NetProbe software looks for patterns that will reveal such so-called "false positives."

The researchers analyzed about 1 million transactions involving 66,000 eBay users to develop graphs - known in statistical circles as bipartite cores - that identify users interacting with unusual frequency. The CMU researchers say that when the transactions between fraudsters and their accomplices are plotted as a graph, they create a pattern that sticks out "like a guiding light."

A paper on the findings of this research is expected to be published early this year, with the software, perhaps, eventually marketed to eBay and/or made available to consumers.

ATR discussed this briefly in a recent podcast and it sheds new light on Alice LaPlante's question, "Is eBay's Feedback System is Fatally Flawed?" that we presented here back in December. Please share your thoughts.

0 Comments:

Post a Comment