How to identify fraudulent reviews?

Question

data mining

Hello everyone
I'm starting to develop a site that will provide feedback (advice). Since the essence of the site is to provide reviews, the question arises how to weed them out (identify fraudulent ones)?
I ask for advice on how or with what help to weed out false / custom reviews.
PS:
Can you tell me how it is in English? Fraud detection doesn't seem to fit.

Answer 1

And purely humanly, is it really impossible to determine that a message like “I orgasm from the site” or “I have been using it for 10 years, my grandmother also took knitting lessons from it” ... Well, how small are you, the order always stinks.

Answer 2

Average weighted frequency. (technically - possible)

Answer 3

In the English-speaking environment, such reviews are called differently:
fake review, fraudulent review, opinion spam, deceptive opinion spam.
There are many articles on these topics.

Answer 4

You have a choice: hands, or hands, or score.