10000 gamma="auto" in SVC · Issue #8361 · scikit-learn/scikit-learn · GitHub
[go: up one dir, main page]
Skip to content

gamma="auto" in SVC #8361

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom
Closed
amueller opened this issue Feb 15, 2017 · 9 comments · Fixed by #10331
Closed

gamma="auto" in SVC #8361

amueller opened this issue Feb 15, 2017 · 9 comments · Fixed by #10331
Labels
Easy Well-defined and straightforward way to resolve Enhancement

Comments

@amueller
Copy link
Member
amueller commented Feb 15, 2017 •
edited
Loading

It's a bit unfortunate that we just changed the parametrization of SVC's gamma (from 0 to "auto"), but I just realized that using 1 / (n_features * X.std()) would be a much much better default than the current 1 / n_features.

Not sure there's a good way to change it though.

I wonder whether we should at some point do a breaking release like matplotlib did, but that might lead to more issues than it solves... (I kind of regret that we didn't change the cv=3 to cv=5 default when doing the model_selection refactoring).
@GaelVaroquaux
Copy link
Member
GaelVaroquaux commented Feb 15, 2017 via email

@jnothman
Copy link
Member
jnothman commented Feb 15, 2017 via email

@GaelVaroquaux
Copy link
Member
GaelVaroquaux commented Feb 16, 2017 via email

@amueller amueller added Easy Well-defined and straightforward way to resolve Enhancement Need Contributor Sprint labels Mar 3, 2017
@amueller
Copy link
Member Author
amueller commented Mar 3, 2017

I think this is worth it, so let's do scale.

@amueller
Copy link
Member Author
amueller commented Mar 3, 2017

For whoever fixes this, it needs a deprecation cycle

@neokt
Copy link
neokt commented Mar 4, 2017

I'm currently working on this.

neokt added a commit to neokt/scikit-learn that referenced this issue Mar 4, 2017
@neokt
gamma=auto in SVC scikit-learn#8361
0551ddd
@neokt neokt mentioned this issue Mar 4, 2017
Closed
@gxyd
Copy link
Contributor
gxyd commented Dec 15, 2017

We had a release 0.19 in which this change wasn't included (we still have gamma='auto' as default). Is this issue in-line to be ready to be present in the next release?

@amueller
Copy link
Member Author

@gxyd this issue is open, so no, as you've seen #8535 is not merged yet.

@gxyd
Copy link
Contributor
gxyd commented Dec 16, 2017

I just realized that using 1 / (n_features * X.std()) would be a much much better default than the current 1 / n_features.

I would be interested to know why is that so (i.e. why 1 / (n_features*X.std() would be a better default)? You can refer me to some material if you have in mind for this statement.

@gxyd gxyd mentioned this issue Dec 16, 2017
Merged
@mgelbart mgelbart mentioned this issue Dec 8, 2018
Closed
@desilinguist desilinguist mentioned this issue Aug 15, 2019
Closed
@glevv glevv mentioned this issue Aug 7, 2022
Open
@glemaitre glemaitre mentioned this issue Oct 25, 2022
Closed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees
No one assigned
Labels
Easy Well-defined and straightforward way to resolve Enhancement
Projects
None yet
Milestone
No milestone
Development

Successfully merging a pull request may close this issue.

[MRG+2] default gamma='auto' in SVC gxyd/scikit-learn
5 participants
@jnothman @GaelVaroquaux @amueller @gxyd @neokt
0