Closed mjskay closed 4 years ago
I'm happy to write a PR for this if you want it. I would propose changing the signature of dist_student_t
from dist_student_t(df, ncp = NULL)
to dist_student_t(df, mu = 0, sigma = 1, ncp = NULL)
and updating the other functions for the student t distribution accordingly.
A PR would be great for this. Thanks!
From an uncertainty vis perspective, having a scaled-and-shifted Student's t dist would be very useful: it's pretty common for both frequentist confidence distributions and Bayesian priors.
As a bonus, if it followed the Stan parameter order (df, location, scale) it would even more useful :).
My humble suggestion would be to modify dist_student_t to use something like (df, mu, sigma, ncp) (assuming not too many people have starting using it yet / or tend to use ncp as a named parameter anyway).