stan-dev / stan-dev/rstanarm

strategies for prior rescaling with cross-validation

Open
#235 1 comment 0 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

Dominant language
R
Stars
401
Forks
136
PR merge metrics
No merged PRs in 30d

Description

@avehtari @dpsimpson @bgoodri here's a summary of our discussion:

Option 1
  • Use MAD instead of SD for rescaling
  • Need to handle case when MAD is 0
Option 2
  • Use SD for everything (including binary variables)
  • Then need to do prior-weighted version of LOO

Our hunch was that option 1 is a good approximation to option 2 but we should do experiments.

Contributor guide

No contributing guide indexed for this repository

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

Research direction

The issue names no files, tests, or entry points. Start by locating the prior-rescaling and cross-validation implementations in the R package, then review how MAD, SD, binary variables, and prior-weighted LOO are currently handled. Done means experiments compare both options and document a selected strategy, including behavior when MAD is zero.

Written by the indexing model from the issue text.

Assessment

Tech stack
r
Domain
data
Issue type
Feature
Difficulty
5/5
Estimated time
Over a week
Activity status
Stale
Clarity
Needs clarification
Newbie friendliness
15/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.