She blinded me with click fraud: Feds fund realtime fraud research
Advertisers aren't the only ones trying to solve the problem of click fraud; the federal government is also interested. Last week, the National Science Foundation issued a $149,923 grant to a graduate student working on the problem as part of the government's Small Business Technology Transfer (abbreviated as "STTR" for some reason) program.
Li Ge, a graduate student at the University of Louisville, has been working on his dissertation since May 2005. His project, a "Collaborative Click Fraud Detection and Prevention System," won a Phase I STTR grant that will cover his work in 2007. Li's approach "draws on data mining techniques for fraud identification using detailed user activities" and works in realtime to detect software clicks. If the Phase I results are good, Li can apply for a Phase II grant of up to $750,000 to continue his work for several more years. Final commercialization of the technology would not be funded by the government, however.
Since Li's system appears to rely on realtime, collaborative processing of clickstream information, it sounds like the kind of research that could produce results that extend far beyond click fraud. As Li himself says in the abstract to the project, the work "will be applied in multiple fields related to online business marketing, user analysis and other fraud identification processes." Companies already use such internal data to learn about customer behavior; running data mining algorithms on a much larger pool of information could produce even more insights into the ways that people use websites (along with new concerns about privacy and tracking).
The STTR was founded by Congress in 1992 as a way to funnel research money to small businesses not often able to compete for big government grants. Participating federal agencies are required to reserve a portion of their yearly research budgets for the program and to award them only to partnerships between a small business and a nonprofit research institution.
The government apparently believes that click fraud is a large enough problem for US businesses that it deserves some federal funding. It's difficult to know exactly how big the problem is, though—as we reported earlier this year, Google believes that the biggest problems lie with the companies that report it, and that the actual issue has been blown out of proportion. Google also touts a professor's report (PDF) that the company commissioned as proof that their click fraud efforts are "reasonable."
But not everyone is convinced. A BusinessWeek cover story from October described a massive worldwide network of "parked sites" filled with nothing but ads and "paid to read" groups that click on them.
Whatever the actual extent of the problem, it is causing some advertisers to have a crisis of confidence in the entire medium of Internet advertising. One dissertation is unlikely to solve the problem, but it's another sign that the government is taking an interest in the problem; the SEC, FBI, and US Postal Inspectors have already stepped up enforcement actions against the bigger click fraud schemes.
:source: Source: http://arstechnica.com/news.ars/post/20061227-8502.html
