I'm female last time I checked, and a paste from my blog makes gender guesser thinks I'm weak male and gender genie says I'm male. Doesn't mean anything, but worth noting.
Grepped this, which ars technica claims is a real chat log. She says just under 200 words on this log, and it comes up as weak male again on Gender Guesser and male on Gender Genie. Her username of `k is removed for the analysis. http://pastebin.com/x69Akp5L
Genre: Formal Female = 509 Male = 971 Difference = 462; 65.6% Verdict: MALE
Computing represents a pretty specialized topic, and most of the sample data with computing-related discussion will be from men. It would be pretty tough for any simple Bayesian analysis to account for this.
Perhaps you should use... fewer ellipses? There are several grammatical patterns females use more than males (and vice versa). I assume that's how the test works, anyway.
Grepped this, which ars technica claims is a real chat log. She says just under 200 words on this log, and it comes up as weak male again on Gender Guesser and male on Gender Genie. Her username of `k is removed for the analysis. http://pastebin.com/x69Akp5L
Here's the article where Ars Technica links to that pastebin http://arstechnica.com/tech-policy/news/2011/02/how-one-secu...