stan-dev / stan-dev/loo

user defined loss/utility functions

Open
#135 0 comments 2 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

feature
Dominant language
R
Stars
157
Forks
38
Avg merge
4d 16h
Merged PRs (30d)
2

Description

Section 4.3 of https://avehtari.github.io/modelselection/diabetes.html demonstrates how to use loo package with other loss/utility function than log score. LOO-R^2 is another example. Both of these are just for point predictions, which is the easiest case.

Make it easier to provide user defined loss/utility function for point predictions. Arguments would be

  • point estimate type given for E_loo (e.g. "mean")
  • user defined function with at least two arguments 1) point predictions and 2) data y

We can later consider expanding to loo/utility functions for predictive distributions.

Contributor guide

Open the contributing guide

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

Start with Section 4.3 of the linked diabetes example and inspect the loo package's existing point-prediction and E_loo entry points. Define how callers provide the point-estimate type and a loss or utility function receiving predictions and y; done means point-prediction functions can use this interface without addressing predictive distributions.

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
Mostly clear
Newbie friendliness
30/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.