Hacker News Reader

Argh, I am so sick of seeing this article. From the study, 16% of all Yelp posts are identified by Yelp as fake, hidden from view, and not counted into a business's average score.

This tells us nothing about the percentage of non-removed Yelp reviews that are fake, because all non-removed reviews are assumed, in the data, to be real.

This is linkbait and a complete misrepresentation of the data.

View on HN

While I agree with you about the title of the article, it is not true at all that "all non-removed reviews are assumed, in the data, to be real." In fact, in our analysis we explicitly allow for the filter misclassifying reviews (that is to say filtering real reviews, or publishing fakes.)

Sorry about my imprecise wording. I wasn't intending to criticize your study, which did acknowledge potential misclassifications (and was very interesting and deserves attention on its own merits).

I'm criticizing MW's interpretation, which assumed that your proxy was very accurate in terms of providing the proportion of fake reviews, but went on to imply that your proxy was highly inaccurate by talking about fake but visible reviews.

Reply on news.ycombinator.com