scikit-learn / scikit-learn/scikit-learn
Investigate SAG/SAGA solver
Nobody has claimed this yet.
- Dominant language
- Python
- Stars
- 67.3k
- Forks
- 27.4k
- Avg merge
- 1d 15h
- Merged PRs (30d)
- 58
Description
Description
The newly introduced tight tests for Ridge in #22910 together with the random seed fixture in #22749 revealed some shortcomings of the sag and saga solver, some are already mentioned in #21305:
It shows some random behavior even with fixed random seed.- The
tolneeds to be set much smaller to receive comparable results with the other solvers ofRidge. - Saga seems more stable on dense data
X, and often needs much more iterations on sparseX. fit_intercept=Trueseems to converge slower.
Ideally, the cause for both issues can be identified and fixed.
Some links for context
https://github.com/scikit-learn/scikit-learn/issues/23014
https://github.com/scikit-learn/scikit-learn/pull/23017
https://github.com/scikit-learn/scikit-learn/pull/23026
https://github.com/scikit-learn/scikit-learn/pull/23152
#23177
#25198
Contributor guide
First steps
- Read the whole issue, then the project's contributing guide.
- Comment on the issue to say you are picking it up — it saves two people doing the same work.
- Fork the repository and make your change on a branch.
- Open a pull request that references the issue number.
Research direction
Start by reading the tight Ridge tests from #22910, the random seed fixture from #22749, and the linked discussions and pull requests, including #21305. Identify the causes of SAG/SAGA convergence and iteration differences, then validate any fix against comparable Ridge results on dense and sparse data with and without an intercept.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- python
- Domain
- machine-learning
- Issue type
- Bug
- Difficulty
- 5/5
- Estimated time
- Over a week
- Activity status
- Stale
- Clarity
- Needs clarification
- Newbie friendliness
- 25/100