Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

change gamma_lcdf to just use gamma_p and the rev-mode derivatives #2763

Closed
spinkney opened this issue Jun 25, 2022 · 2 comments
Closed

change gamma_lcdf to just use gamma_p and the rev-mode derivatives #2763

spinkney opened this issue Jun 25, 2022 · 2 comments

Comments

@spinkney
Copy link
Collaborator

spinkney commented Jun 25, 2022 •
edited

Using gamma_lcdf for values of alpha values approximately > 12 fails estimation. The interesting thing is that we have the Gamma CDF calculated in the gamma_p function. Using this function instead produces the correct output. We should change the gamma_lcdf calculation to just use gamma_p and it's reverse mode derivatives.

I believe #780 fixed the issue for gamma_p (also see the linked issue #525) but the gamma_lcdf was never updated.

The issue encountered is explained in more detail at https://discourse.mc-stan.org/t/minimal-censored-gamma-with-very-bad-performance/27922.
@spinkney
Copy link
Collaborator Author

I think this is buggy enough and hopefully easy enough to get in for the 2.30 standard release.

@andrjohns andrjohns mentioned this issue Jun 26, 2022
Merged
5 tasks
@spinkney
Copy link
Collaborator Author

Closed with #2767

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees
No one assigned
Labels
None yet
Projects
None yet
Milestone
No milestone
Development

No branches or pull requests
1 participant
@spinkney