using yardstick (or other package) for performance calculations

Open
#495 0 comments 1 reaction 0 assignees View on GitHub

Nobody has claimed this yet.

Assessment

Difficulty
5/5
Estimated time
Over a week
Newbie friendliness
25/100
Issue type
Feature
Clarity
Needs clarification
Activity status
Stale
Tech stack
python, r

Research direction

Review the existing loss-function implementation and compare its coverage with yardstick's metrics, direction metadata, multi-metric support, multiclass metrics, censored regression, and user-defined metrics. The issue names no files, tests, entry points, or completion criteria, so confirm the intended scope and R/Python parity requirements with maintainers before coding.

Written by the indexing model from the issue text.

Description

R 🐳

The loss functions included are great but are somewhat limited.

There are a lot of R packages that could expand the types of loss functions (but at the loss of R:python parity).

If you were to use yardstick, for example, the benefits would be:

  • More metrics
  • Data on direction (e.g larger-is-better) for each metrics. You wouldn't have to do 1 - AUC anymore.
  • yardstick can compute multiple metrics at once.
  • Numerous multi-class metrics
  • Metrics for censored regression
  • Extensible for user-defined metrics

The downside to the current system is that you might optimize your model on a set of performance scores and judge feature importance on some other score.

Let use know if we can help or put in a PR.

Dominant language
Python
Stars
1.5k
Forks
172
PR merge metrics
No merged PRs in 30d

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.

More from ModelOriented/DALEX

All issues in ModelOriented/DALEX

Similar issues

More Python issues

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.