8000 Support for Enum-Encoding as in H2o · Issue #11258 · scikit-learn/scikit-learn · GitHub
[go: up one dir, main page]
Skip to content

Support for Enum-Encoding as in H2o #11258

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
dnks23 opened this issue Jun 14, 2018 · 2 comments
Closed

Support for Enum-Encoding as in H2o #11258

dnks23 opened this issue Jun 14, 2018 · 2 comments

Comments

@dnks23
Copy link
dnks23 commented Jun 14, 2018

Hello,

the H2o ML Framework supports an enum-encoding scheme. It would be nice to have this for sklearn as well. As far as I know there are no contributions made to add this for sklearn tree-based-models.

This would be useful to handle categorical features without some curse-of-dimensionality issue (One-Hot) and any kind of ordinality implied. This seems to be a nice approach to find the best split in tree-based-models (e.g. random forest) for categorical features. There is also an implementation of this in LightGBM: Read Section about optimal split for categorical features where there are 2^(k-1) - 1 possible subsets of the k-categorical features for splitting.

Anyone has thoughts about this?
@jnothman
Copy link
Member
jnothman commented Jun 14, 2018 via email

@jnothman jnothman mentioned this issue Aug 29, 2018
Closed
@timbicker timbicker mentioned this issue Sep 8, 2018
Closed
@lorentzenchr
Copy link
Member

For the histogram gradient boosted trees, we now have native categorical support with #18394. You need to pass ordinal encoded columns for this.

@dnks23 I'm closing for now as this seemed to be your main use case. Please feel free to say if you had more in mind.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees
No one assigned
Labels
None yet
Projects
None yet
Milestone
No milestone
Development

No branches or pull requests
3 participants
@jnothman @dnks23 @lorentzenchr
0