Fake online reviews are a pain in the arse: they make interent shopping harder than it already is. But thankfully there are people out there who are developing ways of spotting and blocking rogue five-star reviews, and a new algorithm backed by Google seems to be the most effective yet.
Developed at the University of Illinois, a new spam-ranking algorithm called GSRank — Group Spam Rank to its friends — can filter out the fraudulent reviews. In fact, it “consistently outperforms all existing methods” according to the team behind it.
So, how does it work? It actually uses a system which allows it to spot dodgy reviews by identifying entire groups that are posting dubious comments. It analyses how closely together people post their reviews, when they’re posted, and the similarity of review content in order to establish relationships between spammers and which products they comment on. You can read the full paper here.
From there, it can start working out which reviews, on which products, are spam. Currently there’s no word on when such a system might make an appearance on a live site — but hopefully it will happen soon. [University of Illinois via The Register]
Image by Dusit/Shutterstock