>>59386758>how does someone not type like a maleINTRODUCTION
I have conclusive evidence to suggest Ayamefriend is a female.
I will use two metrics to defend this argument, one is empirically supported by this scientific paper:
https://www.sciencedirect.com/science/article/abs/pii/S174228761730364X?via%3Dihubwhich found specific typing "mannerisms" or "patterns" that can predict gender at a 95% accuracy rate.
And a second metric that is based on female speaking mannerisms, and an assumption that they translate into online text/typing. This is, again, going to be based on sociolinguistic and psycholinguistic research, which has been nicely summed up in this scientific writing article:
https://verilymag.com/2015/03/sex-differences-language-women-uptalk>METHODI scoured through warosu with the keywords "where we at?" and logged the last 10 Ayamefriend posts.
From these posts, I simply counted the incidence of female-like mannerisms as described in the prior two metrics.
These include:
The use of speech fillers like "um", "like", "just", "totally"
The use of overly effusive adjectives/adverbs, "adorable", "cute", "amazing", "beautiful"
Keystrokes that maximize female information gain (IG; reduces entropy/variance) in gender classification: N-A (0.0897), M-O (0.0815), K-A (0.0706), R-I (0.0647), M-A (0.0612). These are set for a one sentence prompt (length: 63 chars), so extrapolate for an average Ayamefriend post (length: 1810 characters; 1810/63 = 28.73, so all IG values are divided by 28.73).
>RESULTSSOURCE (listed 1 to 10):
>>3148663#>>3145497 #>>3139367 #>>3119255 #>>3117272 #>>3114751 #>>3113048 #>>3110686 #>>3102804 #>SPEECH FILLERS"Like" count: 3, avg 0.3
"um" count: 0
"just" count: 9, avg 0.9
"totally" count: 0, avg 0
"adorable" count: 9, avg 0.9
"cute" count: 4, avg 0.4
"amazing" count: 5, avg 0.5
"beautiful" count: 3, avg 0.3
"fantastic" count: 2, avg 0.2
"sweet" count: 8, avg 0.8
"lovely" count: 2, avg 0.2
Assuming a standard phrase structure sentence as described by linguist Noam Chomsky in his 1957 book with an average ratio of 1 adverb/adjective/modiffier to 3 words (1:2.66, specifically).
Average total word count of all messages: 101 x 10 - 1010, then divided by the ratio (1010/2.66) = 379 expected sentence adhesivestotal "girly" sentence adhesives: 49, meaning 13% of Ayamefriend's sentence modifiers are considered "girly". That's a pretty high number for something that should be near zero.
>KEYSTROKES"N-A": 44
"M-O": 74
"K-A": 10
"R-I": 87
"M-A": 89
So, (44 x 0.0815/28.73) + (74 x 0.0815/28.73) + (10 x 0.0706/28.73) + (87 x 0.0647/28.73) + (89 x 0.0612/28.73) = 0.74482.
75% chance Ayamefriend is a female by keystroke alone. Then account for the 95% success rate in modeling, (0.95 x 0.74882) = 0.7068.
>CONCLUSIONAyamefriend is very likely a female. There is a 71% chance she is a female by keystrokes alone, and that chance is further increased (though we are unsure by how much; two different metrics) when accounting for her speech adhesives and filler words. Note that I excluded many other "girly" speech fillers and modifiers like "comfy" and "cute". Alas.
