KEMBAR78
DOC: fix convergence warning in iterative example by deniederhut · Pull Request #14330 · scikit-learn/scikit-learn · GitHub
Skip to content

Conversation

@deniederhut
Copy link
Contributor

Reference Issues/PRs

Resolves on check box from #14117

What does this implement/fix? Explain your changes.

The estimator fails to converge reliably even with 1e4
iterations. We can keep the default max_iter and avoid
convergence warnings by increasing the tolerance.

Any other comments?

Before change:

Screenshot 2019-07-13 11 35 49

After change:

Screenshot 2019-07-13 11 36 05

The estimator fails to converge reliably even with 1e4
iterations. We can keep the default max_iter and avoid
convergence warnings increasing the tolerance.
@amueller
Copy link
Member

That seems a bit weird. Do you know which estimator this is for?

@amueller
Copy link
Member

This is really strange, the estimator is not converging at all. This looks like a bug to me. I opened #14338 and I think we should investigate the actual issue.

@deniederhut
Copy link
Contributor Author

😬

@deniederhut
Copy link
Contributor Author

Is this good to go? Or is the idea that we want to wait for the fixes coming to the IterativeImputer first?

@amueller
Copy link
Member

@deniederhut I would argue that it's not good to try and hide the warning and we should fix the underlying bug.

@deniederhut
Copy link
Contributor Author

Cool, then let's close this in favor of whatever PR closes #13773

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants