Skip to content
a prototype greta extension for specifying priors over *parametric* functions
Branch: master
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Type Name Latest commit message Commit time
Failed to load latest commit information.


Build Status Codecov test coverage

It’s common in Bayesian statistical analyses to have prior information on summary statistics of a model rather than on the model parameters themselves. For example, you may have prior information about the probability predicted by a logistic regression under some set of covariates, or the timing of peaks in a dynamical model. In these situations it can be difficult to determine the priors on the model parameters that correspond to these priors on summary statistics. greta.funprior lets you define priors on these summary statistics by considering them as priors over parametric functions, and provides a simple interface for fitting models.

This is an early prototype based on the greta.template template package. The functionality hasn’t been added yet.

This package has a Contributor Code of Conduct. By contributing to this project, you agree to abide by its terms.

to do list

  • Pick a package name (preferably with ‘greta.’ at the beginning). Update the package name in:
    • the ‘Package’ field in DESCRIPTION
    • the library() and test_check() calls in tests/testthat.R
    • in the @name documentation field in R/package.R
    • at the top of README.Rmd
    • the repo name (if it changed since you made it)
  • Come up with a helpful package title. Add it to:
    • the ‘Title’ field in DESCRIPTION
    • the @title documentation field in R/package.R
    • at the top of your GitHub repo
  • Fill in the the author details in DESCRIPTION
  • Update the ‘URL’ and ‘BugReports’ fields in DESCRIPTION to point to your repo
  • Decide what sort of license you want to use for your package (you are completely free to change the CC0 license in the template). See or for help choosing.
  • Write short paragraph describing the package. Copy it to:
    • the ‘Description’ field of DESCRIPTION
    • the @description documentation field of R/package.R
  • Log in to travis (e.g. with your GitHub credentials) and turn on tracking of your new repo
  • edit the travis and codecov badges in README.Rmd to point to your package
  • Write a simple example introducing the package and add it to R/package.R
  • Update README.Rmd to sell to people with a sales pitch and maybe an example that creates a figure, to get people excited about your package.
  • Start adding functions, documentation and examples to new R files in R folder
  • Write some unit tests for these functions in the tests/testthat folder
  • Delete the example function and test files R/square.R and tests/testthat/test-square.R
You can’t perform that action at this time.