Men talk about stuff, women talk about relationships
A late night at the office, and after a while the brain seizes up and fizzles out like an overheated car engine and one has to take a break while listening to the drone of the cleaning guy's vacuum. Tonight's diversion is The Gender Genie, an online application parses a block of text and guesses whether the author is male or female.
Naturally I plugged in some of my own writings, and then I plugged in some posts from Jenny's weblog as well. The Gender Genie did well. Occasionally I write like a girl, but usually not, and Jenny is most definitely a female, which both she and Adam will be glad to know since she gave birth to a child recently.
The algorithm is discussed in both the NYTimes and Nature. From the Nature article:
The program's success seems to confirm the stereotypical perception of differences in male and female language use. Crudely put, men talk more about objects, and women more about relationships.
Female writers use more pronouns (I, you, she, their, myself), say the program's developers, Moshe Koppel of Bar-Ilan University in Ramat Gan, Israel, and colleagues. Males prefer words that identify or determine nouns (a, the, that) and words that quantify them (one, two, more).
Supposedly the algorithm should be correct about 80% of the time, though the online app, which asks its users if it was right or not, is currently performing below 50%. Who knows if that's true or just the result of jealous users. Supposedly the algorithm is even better at detecting differences between non-fiction and fiction, achieving 98% accuracy.
I'd be interested to compare the algorithm's success rate with simple human judgment. I wonder if humans have an innate ability to detect an author's sex from the text itself the same way humans have a very perceptive intuition about other people's emotions.
And yes, it should be called the Sex Genie, because only pronouns have gender, people have sex. But perhaps calling this software app the Sex Genie might cause undue excitement among some Internet users or might infringe upon the copyright of some kinky toy.
Naturally I plugged in some of my own writings, and then I plugged in some posts from Jenny's weblog as well. The Gender Genie did well. Occasionally I write like a girl, but usually not, and Jenny is most definitely a female, which both she and Adam will be glad to know since she gave birth to a child recently.
The algorithm is discussed in both the NYTimes and Nature. From the Nature article:
The program's success seems to confirm the stereotypical perception of differences in male and female language use. Crudely put, men talk more about objects, and women more about relationships.
Female writers use more pronouns (I, you, she, their, myself), say the program's developers, Moshe Koppel of Bar-Ilan University in Ramat Gan, Israel, and colleagues. Males prefer words that identify or determine nouns (a, the, that) and words that quantify them (one, two, more).
Supposedly the algorithm should be correct about 80% of the time, though the online app, which asks its users if it was right or not, is currently performing below 50%. Who knows if that's true or just the result of jealous users. Supposedly the algorithm is even better at detecting differences between non-fiction and fiction, achieving 98% accuracy.
I'd be interested to compare the algorithm's success rate with simple human judgment. I wonder if humans have an innate ability to detect an author's sex from the text itself the same way humans have a very perceptive intuition about other people's emotions.
And yes, it should be called the Sex Genie, because only pronouns have gender, people have sex. But perhaps calling this software app the Sex Genie might cause undue excitement among some Internet users or might infringe upon the copyright of some kinky toy.