View Full Version : Gender Prediction -- Can You Fool It?
I find this fascinating -- using an algorithm, this software predicts from samples of your writing whether you are male or female. I can fool it when I deliberately try to do so, but I can't trick it at all when I paste in chunks of text I wrote in letters or posts. Pretty slick. Give it a try!
Gender Writing Test (http://www.bookblog.net/gender/genie.html)
Big_John
12-28-2006, 15:52
a 396-word passage from melville's Bartleby the Scrivener threw it. :undecided:
edit: unless........ :shifty:
I would imagine that the software is "tuned" for modern English. Tossing in 19th-century prose (and Melville borders on the color purple -- the guy used what were considered anachronisms by the standards of his own time) will probably confuse it.
I think it's more fair to feed it your own writing, and see how it does.
KukriKhan
12-28-2006, 16:08
Pasted in a small bit of frogbeastegg's MeadHall work. Verdict:
Words: 137
(NOTE: The genie works best on texts of more than 500 words.)
Female Score: 204
Male Score: 101
The Gender Genie thinks the author of this passage is: female! Good Job!
But, pasting in my post from the frontroom beer thread, it got:
Words: 91
(NOTE: The genie works best on texts of more than 500 words.)
Female Score: 83
Male Score: 60
The Gender Genie thinks the author of this passage is: female!
:laugh4:
In fairness, those pieces weren't long enough, probably.
Interesting division of Masculine v Feminine words they're using..
Big_John
12-28-2006, 16:22
a 1500 word closed memo i wrote a couple months ago threw it also. :undecided:
unless........ :shifty:
Rodion Romanovich
12-28-2006, 16:39
Apparently Shakespeare (Hamlet) gets 1838 male points and 2666 female points, and Shakespeare is thereby officially a woman (wasn't there a conspiracy theory saying that Shakespeare's works were really written by a woman? :grin: )! :wink:
Sasaki Kojiro
12-28-2006, 17:04
Female Score: 647
Male Score: 817
The Gender Genie thinks the author of this passage is: male!
Nifty.
Somebody Else
12-28-2006, 17:56
That speech I submitted in the Mead Hall a little while ago...
Female Score: 2154
Male Score: 2526
The Gender Genie thinks the author of this passage is: male!
Reassuring I suppose...
*edit*
Just stuck one of Cicero's speeches through the mill... apparently he's female. Not so reassuring, I was supposed to be mimicking his style!
A 400 word hardboiled piece I wrote was called female. I would like to understand the logic behind it.
edyzmedieval
12-28-2006, 20:34
Words: 747
(NOTE: The genie works best on texts of more than 500 words.)
Female Score: 864
Male Score: 743
My translation of a bit of the Aeneid fooled it totally
Words: 1197
(NOTE: The genie works best on texts of more than 500 words.)
Female Score: 2389
Male Score: 1216
The Gender Genie thinks the author of this passage is: female!
----------------------------------------------------------
A report I did for 12th Century Glory got:
Words: 805
(NOTE: The genie works best on texts of more than 500 words.)
Female Score: 1107
Male Score: 1111
The Gender Genie thinks the author of this passage is: male!
----------------------------------------------------------------
One of my darker short stories, SELB, for short, got:
Words: 1043
(NOTE: The genie works best on texts of more than 500 words.)
Female Score: 1429
Male Score: 1075
The Gender Genie thinks the author of this passage is: female!
---------------------
On another short story (humour)
Words: 689
(NOTE: The genie works best on texts of more than 500 words.)
Female Score: 1241
Male Score: 780
The Gender Genie thinks the author of this passage is: female!
------------------------------------
Oh dear.
Alexander the Pretty Good
12-28-2006, 21:14
Amusing, but is the entire algorithm the tables on the bottom? Seems overly simplistic to me. Even though it has my writing correct...
Unless... :shifty:
Banquo's Ghost
12-28-2006, 21:42
It appears the first chapter of my story posted in the Mead Hall makes me female.
:book:
Big_John
12-28-2006, 21:55
It appears my the first chapter of my story posted in the Mead Hall makes me female.
:book:
....... unless............. :shifty:
Rodion Romanovich
12-29-2006, 00:01
Cool! My post about meat (https://forums.totalwar.org/vb/showthread.php?t=76293) was the most masculine text I've found so far!
Female Score: 359
Male Score: 712
The Gender Genie thinks the author of this passage is: male!
This thing thinks I am Female, through and Through!
Everything I get has about this composition:
Words: 527
(NOTE: The genie works best on texts of more than 500 words.)
Female Score: 709
Male Score: 634
The Gender Genie thinks the author of this passage is: female!
http://www.bookblog.net/gender/female1.gif
I Most certainly am not female!!
Unless... :shifty: (sorry for using it, but It kills me!)
Avicenna
12-29-2006, 06:01
Words: 3626
(NOTE: The genie works best on texts of more than 500 words.)
Female Score: 2865
Male Score: 5773
Marshal Murat
12-29-2006, 06:30
So is this a measure of your masculinity?
A Short Play, THE MASCULINE
"I've got 5039!"
"Oh yea?"
"Yea!"
"I've got 6493!"
"I'm, so sorry to have bothered you, sir, I mean, pimp, I mean, uh, teach me your ways!"
"Okay fine, just use your and he alot. Now get off my leg, its freaking me out!"
CountArach
12-29-2006, 07:30
Words: 863
(NOTE: The genie works best on texts of more than 500 words.)
Female Score: 815
Male Score: 1080
So it got it right. I used some story I wrote here:
http://z2.invisionfree.com/herdstone/index.php?showtopic=8282
Stupid.
I copied from a friend's(female) Myspace blog and it though it was male?!?
Its completely foolish. It is just a random ensamble of stereotypical words used by both sexes.
Marshal Murat
12-30-2006, 06:31
Actually no, men are more predisposed to use certain words while women are more likely to use others.
Linky to the Israeli site/research please?
While the algorithim to the whole deal is very basic, if there was more variables or something, then it might work.
Linky to the Israeli site/research please?
I haven't gotten the link to work yet, but this PDF (http://www.cs.biu.ac.il/~koppel/papers/male-female-text-final.pdf) is supposed to be the paper ...
Actually no, men are more predisposed to use certain words while women are more likely to use others.
Linky to the Israeli site/research please?
While the algorithim to the whole deal is very basic, if there was more variables or something, then it might work.
keyword: might
scotchedpommes
12-30-2006, 08:03
A similar genie began floating around blogs a number of years ago. Would say it's far from being absolute nonsense. Would also imagine numerous factors could contribute to development of alternative word choice. [without having looked at the algorithm or paper, neither of which are working for me] If I remember rightly, the previous test focused on use of specific contractions or generalisation, and references to self. Based on individual habits and inevitable variations it's easy to see how we can all become a bit 'confused'.
{Alright, having actually taken the test, it does admittedly seem even more simplistic than the last. Slightly disappointing in its limitations.}
vBulletin® v3.7.1, Copyright ©2000-2025, Jelsoft Enterprises Ltd.