Skip to content
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

docs: Use logistic regression instead of SVC on the Iris dataset #1018

Closed
sylvaincom opened this issue Dec 24, 2024 · 1 comment · Fixed by #1087
Closed

docs: Use logistic regression instead of SVC on the Iris dataset #1018

sylvaincom opened this issue Dec 24, 2024 · 1 comment · Fixed by #1087
Assignees
@sylvaincom
Labels
documentation Improvements or additions to documentation

Comments

@sylvaincom
Copy link
Contributor

sylvaincom commented Dec 24, 2024 •
edited
Loading

Which part of the documentation needs improvement?

README, quick start example

Describe the problem found in the documentation

Currently, on the Iris dataset, for classification, we use SVC.

Suggested improvement

@glemaitre suggests using logistic regression.

Additional context

Related to #1004.
Waiting on this PR to be merged: #1009
@sylvaincom sylvaincom added documentation Improvements or additions to documentation needs-triage This has been recently submitted and needs attention labels Dec 24, 2024
@sylvaincom sylvaincom self-assigned this Dec 24, 2024
@sylvaincom sylvaincom mentioned this issue Dec 24, 2024
Merged
@glemaitre
Copy link
Member

Some background here: SVC does not really scale with the number of samples. It was useful at the time that kernels were popular. But nowadays, I would even advocate with a kernel approximation and a logistic regression to achieve the same thing and it will scale.

So as a general rule, I think that we should really either show:

  • LogisticRegression when it comes to classification with linear model
  • HistGradientBoosting in regression and classification when we want to show the direct state of the art models.

Then, we can use any other classifier or regressor but it means that we want to show a specific feature of this particular model.

@tuscland tuscland removed the needs-triage This has been recently submitted and needs attention label Jan 3, 2025
@sylvaincom sylvaincom mentioned this issue Jan 10, 2025
Merged
@sylvaincom sylvaincom linked a pull request Jan 10, 2025 that will close this issue
docs: Use logistic regression instead of SVC on the Iris dataset #1087
Merged
thomass-dev pushed a commit that referenced this issue Jan 13, 2025
@sylvaincom @thomass-dev
docs: Use logistic regression instead of SVC on the Iris dataset (#1087)
d25fd2d 
Fix #1018 

In quick start example, do not use SVC on the Iris dataset, but rather
logistic regression
rouk1 pushed a commit that referenced this issue Jan 14, 2025
@sylvaincom @rouk1
docs: Use logistic regression instead of SVC on the Iris dataset (#1087)
9ce24f6 
Fix #1018 

In quick start example, do not use SVC on the Iris dataset, but rather
logistic regression
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees

@sylvaincom sylvaincom

Labels
documentation Improvements or additions to documentation
Projects
None yet
Milestone
No milestone
Development

Successfully merging a pull request may close this issue.

docs: Use logistic regression instead of SVC on the Iris dataset probabl-ai/skore
3 participants
@tuscland @glemaitre @sylvaincom