Suppose a person (John Doe) uses different user names (handles) in different circumstances (for simplicity, let us limit the conversation to a single social network, e.g., twitter). E.g., he

  • posts pictures of kittens as john_doe,
  • posts political rants as deplorable_john,
  • shares technology news as super_doe
  • &c (and each screen_name has a different id, so no short cut).

How would you go about detecting that all these accounts are used by the same individual?

  • SNA: look at who they follow (and maybe who follows them): if these sets are sufficiently similar, it's a good signal.

  • NLP: compare the posted texts and try to detect that they have been written by the same author.

  • Meta data: see if the accounts are usually active from the same geo location at the same time.

What other approaches have been / could be tried?

Any published work?

  • $\begingroup$ You'll find much published work by searching for "author(ship) recognition/attribution/verification/classification". $\endgroup$ – Emre Jan 10 '17 at 18:01
  • $\begingroup$ You can also search for "computational stylometry" - en.wikipedia.org/wiki/Stylometry $\endgroup$ – Brandon Loudermilk Jan 12 '17 at 15:15

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.