Adding `sample_weight`support to `DummyRegresor` #3420

Closed
arjoly opened this Issue Jul 18, 2014 · 8 comments

Projects

None yet

3 participants

@arjoly
Member
arjoly commented Jul 18, 2014

DummyRegressor could easily support sample_weight.

The main challenge is to implement a weighted quantile median function. For reference,

@arjoly
Member
arjoly commented Sep 22, 2014

The weighted percentile is already implemented in a private function of sklearn.ensemble.gradient_boosting. (Thanks to @pprett )

@shibi-raj

Hi, I'm new to sklearn and interested in working on this enhancement. I'll take a look at the references and see what I can do.

@arjoly
Member
arjoly commented Oct 7, 2014

Thanks @shibi-raj !

@arjoly
Member
arjoly commented Oct 16, 2014

@shibi-raj I finally needed this for a project. So I made a quick pull request at #3779.

@MechCoder
Member

Arnaud, I would like to know the place in your project where a Dummy Regressor is used, just to learn.

@arjoly
Member
arjoly commented Oct 20, 2014

A regression dataset where some samples are individuals that are present multiple times.

@MechCoder MechCoder closed this Nov 4, 2014
@MechCoder
Member

@arjoly If you reference something like Fixes in the PR message or a commit, this will close automatically once the PR is merged.

Just a tip :)

@arjoly
Member
arjoly commented Nov 4, 2014

Thanks!

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