How to write an algorithm that filters out fake names?

Question

As you probably know, Google+ and Facebook prohibit the user from creating an account with a “fake” name.

What is it worth? How can you suspect that the name is unrealistic?

Why is Paul Smith bigger than John John?

Can there be statistics on existing usernames?

Edit (after this question has been closed): I think this question is constructive - what will you do if you are a google software developer who needs to implement the "fake name" function ...

Answer 1

This is not a matter of programming really IMHO. Because they are global services, G + probably uses heuristics based on the sheer amount of data that Google has collected on what could be a name.

Answer 2

These are almost all heuristics. They have a blacklist of names or names that, in their opinion, SOUND are faked - that is, tied to current events, geography, holidays, etc. ...

Answer 3

There is no real way to determine if a name is fake or real. Facebook and Google+ usually do not delete / disable accounts based on a name like John John , which may be a valid name. Google+ began to disconnect accounts with ".", "*" Or any other special characters as the only last name or first name ( link ). There are other cases when the account is disabled, because they look like celebrities ( link ).

Alternatively, they may prohibit heuristic-based accounts, but I don't think this is mandatory. You can create a fake account with the name John John or Iam Awesome on facebook, and you won’t be banned (Source: My friend has had a BS account on Facebook for more than a year with an eccentric name, and his account is still good.)