-1
$\begingroup$

Does anyone know of a trained net that can classify gender from username from any website? It would be nice if there is something in Wolfram Neural Net Repository or something that can be easily imported and used in Wolfram Framework.

I am performing data analysis on a dataset from a website but it lacks the gender field. It however has the username. I would like to add a predicted gender field so I can preform some analysis. Please if someone know of such a net I want to know.

If it helps I also have user's country of origin which might help disambiguate some username's gender.

$\endgroup$
  • $\begingroup$ Possible duplicate of stackoverflow.com/questions/818203/… $\endgroup$ – Roman May 15 at 13:40
  • 2
    $\begingroup$ Seeing your username, I wonder how you came to the idea of classifying users' gender by username... $\endgroup$ – Henrik Schumacher May 15 at 14:14
  • 2
    $\begingroup$ his/her username is related to her/his reputation $\endgroup$ – J42161217 May 15 at 14:48
  • 1
    $\begingroup$ @HenrikSchumacher I am not classifying user name from this website. Also I understand some will still remain ambiguous but that is ok where it is impossible. $\endgroup$ – user13892 May 15 at 14:55
  • $\begingroup$ @J42161217 Hah! Indeed! =D $\endgroup$ – Henrik Schumacher May 15 at 15:02
1
$\begingroup$
Classify["NameGender", "Alexey"]

"Male"

$\endgroup$
  • 2
    $\begingroup$ I wouldn't trust this classifier: it already bugs out on Classify["NameGender", "Andrea"] as Mathematica thinks it's a female name and doesn't admit it's highly ambiguous. $\endgroup$ – Roman May 15 at 14:57
  • 1
    $\begingroup$ I tried this. This does not appear to be a real classifier, as it gives Probability 50-50 or Indeterminate for all cases which it doesn't "explicitly" know about. It looks like a basic name lookup rather than a machine learning classifier. It has no intelligence to infer the gender. $\endgroup$ – user13892 May 15 at 15:01
  • 2
    $\begingroup$ @Roman There's Classify["NameGender", "Andrea", "Probabilities"], but in this case it thinks it's 99% female. $\endgroup$ – kirma May 15 at 15:34
  • 1
    $\begingroup$ Thanks @kirma that's a useful indicator. Yes 99% female is a bit suspicious for a name that literally means "man" :) $\endgroup$ – Roman May 15 at 15:41
  • $\begingroup$ Apart from 50-50 probabilities, it's very rare to see this classifier taking an uncertain stance, that is more than 1% for the losing gender. Out of several unisex English names only couple rise above 2%, Riley being an extreme case at 55-45. $\endgroup$ – kirma May 15 at 15:43

Your Answer

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

Not the answer you're looking for? Browse other questions tagged or ask your own question.