A link to a Mashable article that links to a CNet article that has a dead link to a Cornell PDF file of a Powerpoint presentation at conference? I’m lost.
I was hoping for peer reviewed peer review too.
I’m not sure why bloggers link to down stream articles so often rather than following them upstream and giving a H/T to the downstream sources? Is it that hard to follow links? It’s kind of frustrating to click upstream and discover a dead end.
Here’s the actual paper: http://myleott.com/neg_opspam_NAACL2013.pdf
It is an interesting paper, but it is predicated on the idea that the Mechanical Turk people they hired to write fake negative reviews for hotels are representative of fake reviews of hotels. What they have really figured out is how to detect people they’ve hired through Mechanical Turk to write fake, negative hotel reviews. Certainly useful, and a great start, but still limited by the approach they had to take to be sure the fake reviews were fake.
There’s also more modern astro turfing, fake reviews that aren’t completely positive in order to seem plausible.
And reviews don’t have to be fake to be fake - that is, heavy marketing or incentives by the seller of a product or service can create “fake” reviews by people who actually did buy or use the product. Just look at Mediabridge, with over 1500 5 star reviews for its obscure, rebadged router. That’s more 5 star reviews than the best selling, heavily marketed, well received Bose Q15 headphones have. Yet all of those 5 star reviews could be “real” by actual purchasers of the product, but solicited specifically by the company.
This topic was automatically closed after 5 days. New replies are no longer allowed.