output a value between 0 (definitely cis) and 1 (definitely trans)
I think the real predictor here would be name and age. I think trans people generally tend to pick names that are considered cool at the time they choose the name, which means they'll tend to have similar names to babies who are being named around the same time. So a thirty year old named "Eli" is more likely to be trans because that name peaked in popularity in the mid 2010s
maybe instead of age as the 2nd variable, year of naming might make more sense. because i think you would need some cis entries so that you have a good spread of sample data, and you can figure out you're trans at any age - i didn't solve that puzzle until I was like 34, but I know plenty of people who knew in high school, or there's even a couple of folks figuring this stuff out at 40+ or 50+