Hi folks, (01)
Looking into the Nature article, it said:
"Crudely put, men talk more about objects, and
women more about relationships."
as a summary of the algorithm. (02)
I wonder if the misidentification of gender by the algorithm may also
be a hint toward the systemic/holistic nature of the writing. I
imagine that somebody shifting away from mechanistic mental models
toward systemic ones would be more likely to focus on relationships
over objects. The algorithm may become more accurate at guessing
gender with further training and refinement, and it may also point to
ways at guessing mental models. (03)
Take care,
Miroslav (04)
--- Eugene Eric Kim <eekim@blueoxen.org> wrote:
> I recently ran across the Koppel-Argamon algorithm for guessing an
> author's gender by analyzing text:
>
> http://www.nytimes.com/2003/08/10/magazine/10wwln-test.html
>
> The algorithm was incredibly simple, so I decided to whip up a Perl
> module, and play with it a bit. I report my results on my blog:
>
> http://www.eekim.com/blog/tech/perl/gender.html
>
> The brief summary: It doesn't work that well. :-)
>
> -Eugene
>
> --
> This message is archived at:
>
>
http://collab.blueoxen.net/forums/cgi-bin/mesg.cgi?a=tools-yak&i=20030822230424.GA32135@douge.blueoxen.net
> (05)
--
This message is archived at: (06)
http://collab.blueoxen.net/forums/cgi-bin/mesg.cgi?a=tools-yak&i=20030823091156.52483.qmail@web10004.mail.yahoo.com (07)
|