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

add option for using posterior predictive in cross-validation #2517

Closed
wants to merge 1 commit into from
Closed

add option for using posterior predictive in cross-validation #2517

wants to merge 1 commit into from

Commits on Jun 13, 2024

  1. add option for using posterior predictive in cross-validation (facebo… 

    …ok#2517)

Summary:
Pull Request resolved: facebook#2517

see title. This change is particularly important for model selection using the NLL if we have noisy observations. Using the posterior over the true function and not the noisy observations gives quite misleading results about model calibration.

I also think that predicted vs actual plots from LOOCV are insightful when using the posterior predictive when the observations are noisy. We may want to consider adding observation_noise to `predict`, but we can do that in a follow-up.

Reviewed By: Balandat

Differential Revision: D58227612
    @sdaulton @facebook-github-bot
    sdaulton authored and facebook-github-bot committed Jun 13, 2024
    Configuration menu
    Copy the full SHA
    9f8e4bc View commit details
    Browse the repository at this point in the history