Closed emitanaka closed 2 years ago
I can update the arguments + deprecation notice in a pull request if you want
What do you think about aliasing the arguments, so that either mu
or mean
and sigma
or sd
can be used?
Possible nitpick: for student t, sigma is not the standard deviation unless df = Inf, no? Perhaps a more "semantic" alias for sigma in student t would be "scale"?
I think that's reason enough to not have mean
and sd
for student t distributions. The motivation for 'how stats does it' doesn't apply for this distribution anyway. Thanks for nitpicking :+1:
I reckon you should have the argument names as
mean
andsd
and deprecatemu
andsigma
fordist_normal
anddist_student_t
. This is consistent with how you do other distributions, e.g. "shape" and "rate" in gamma distribution + how thestats::dnorm
works too.