Cornell Develops Algorithm for Detecting Fake Online Reviews



+ Add a Comment


I do a lot of online purchasing and I spend a lot of time reading reviews from various websites to get a good idea if a product is worth it.  I've seen reviews where people just copy and pasted what someone else wrote about a product, the person said a product was so great/awful in one sentence without explaining their decision.  Then you always have that one idiot that has a bad experience with the company but leaves a bad review on the product.  I've also seen sites that allowed people to post two exact same reviews for the same product.  Some sites review people's reviews first before posting it publicly which could be a good or a bad thing.  Some sites are able to delete reviews off their site.  I'm sure Sony won't be happy if you visited their website and left a really bad review on their products.  I can imagine a negative review on one of your products on your own website would be bad for sales.  A lot of manufacturers will pay a website to feature their product or put their product at the top of the search list, so I wont' be surprise if they paid off websites to remove bad reviews or to allow posting of bogus reviews.

I think Cornell will have to work a lot harder on their algorithm.  I'm sure someone is going to find away to write reviews to fool the algorithm and just copy & paste template reviews on various products.  One thing I've learned in my lifetime is that when there is a will, there is a way.



So basically, it's a system for detecting the use of euphemisms...


Anyone else thinking about how this could be used in the political spectrum?



Double posts for double the fun.

Log in to MaximumPC directly or log in using Facebook

Forgot your username or password?
Click here for help.

Login with Facebook
Log in using Facebook to share comments and articles easily with your Facebook feed.