Skip to content

Instantly share code, notes, and snippets.

@bdatko

bdatko/carlin_gist.ipynb

Last active February 23, 2022 01:45
Show Gist options
  • Star 1 You must be signed in to star a gist
  • Fork 1 You must be signed in to fork a gist
  • Save bdatko/7bd1b539cda3ec4ee3552e22221443ff to your computer and use it in GitHub Desktop.
Save bdatko/7bd1b539cda3ec4ee3552e22221443ff to your computer and use it in GitHub Desktop.
Download ZIP
Carlin et al. Smoking Longitudinal Binary Mixture Model
Display the source blob
Display the rendered blob
Raw
carlin_gist.ipynb
Sorry, something went wrong. Reload?
Sorry, we cannot display this file.
Sorry, this file is invalid so it cannot be displayed.
@bdatko
Copy link
Author

bdatko commented Feb 10, 2022

Trying to reproduce the fit from A case study on the choice, interpretation and checking of multilevel models for longitudinal binary outcomes. I obtained the data from ARM book data, smoking_pub.dat Accessed Feb 10, 2022 If you are looking at the first version I am still have some issues with the mixture model, and I am posting on the NumPyro discourse for some help.

@bdatko
Copy link
Author

bdatko commented Feb 11, 2022

I updated the notebook to reflect the discussion on NumPyro forum.

  • For the subject level SD we need to re-parameterize the model, it was suggested to use the handlers.reparam
  • The Bernoulli mixture model is very similar to ZeroInflatedDistribution

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment