Rstanarm Developer Notes

This note is designed to help a developer contribute to rstanarm, which requires an understanding of how the various components of the package fit together.

Some things to keep in mind going forward:

  • rstanarm objects are S3 objects of class 'stanreg'.
  • rstan objects are S4 objects of class 'stanfit'.

Before working on including a model into rstanarm you should get the general form of the model(s) working using rstan. Include generated quantities such as posterior predictions, log_lik, and the mean posterior predictive distribution (mean_PPD). This will be useful in debugging the same model that you’ll implement in rstanarm.