4
$\begingroup$

Suppose users in a system like a social network are described by a number of tags. The number of tags can be assumed to be less than 10.

Example

John: funny musician geek professor
Peter: skinny tall soccer manager geek
Amelia: musician feminist power soccer
Kate: jazz drums night bar cinema art
Richard: painting books art cinema geek

Let → mean "follows". We have that:

John → Richard
John → Kate
John → Amelia
Kate → Richard
Amelia → Peter
Peter → John

or writing the tags:

funny musician geek professor → painting books art cinema geek
funny musician geek professor → jazz drums night bar cinema art
funny musician geek professor → musician feminist power soccer
jazz drums night bar cinema art → painting books art cinema geek
musician feminist power soccer → skinny tall soccer manager geek
skinny tall soccer manager geek → funny musician geek professor

Is there a comprehensive logic/theory/model to predict for a new user who he is likely to follow based on his tags? If so, which one? And how? This smells like conditioned probabilities with some assumptions in the mix....

You can use as example James.

James: musician geek police soccer

Who is the person James is more likely to follow?

$\endgroup$
  • $\begingroup$ Smells like a graph theory problem to me. Here is one research paper dealing with something similar, maybe you will find some useful results and/or references therein. $\endgroup$ – Richard Hardy Dec 8 '14 at 14:34

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.