Is there a plan to add a 4 parameters beta disterbution?

Development
1

Hi Team,
I wanted to know if there is a plan to add a built in option for a 4 parameter beta distribution?
I currently implement a lazy “work-around” using pm.Deterministic
Using parameters “a” and “b” regularly and then multiply by “c” to get the width and add “d” for the distance from 0.

Thanks!
Nachshon

2

No plans for that. Why not just create a helper function to do that for you?

def scaled_beta(name, alpha, beta, c, d):
  beta = pm.Beta(f"{name}_raw", alpha, beta)
  return pm.Deterministic(name, beta * c + d)
1 Like
3

Thanks for the response.
This is basically what I did (w/o the function)

beta = pm.Beta("beta", a, b)
β = pm.Deterministic("β", beta * c + d)

I wondered if there was a more elegant way.
Thanks!

4

I think that’s as simple as it gets. Nothing wrong with it

3 Likes