search

scikit-learn-contrib / forest-confidence-interval

Confidence intervals for scikit-learn forest algorithms
http://contrib.scikit-learn.org/forest-confidence-interval/
MIT License
284 stars 48 forks source link

utilize n_samples_bootstrap when generating sample indices #85

Closed emptymalei closed 4 years ago

emptymalei commented 4 years ago

This is a breaking change. It requires scikit learn version >= 0.22. Backward compatibility will be done if it is required.

scikit learn has introduced a new parameter n_samples_bootstrap when generating sample indices in v0.22:

https://github.com/scikit-learn/scikit-learn/blob/07ceb6e2552a8af5e35dd652d24989b6a69dd7c7/sklearn/ensemble/_forest.py#L115-L122

The new parameter is generated by the function _get_n_samples_bootstrap:

https://github.com/scikit-learn/scikit-learn/blob/07ceb6e2552a8af5e35dd652d24989b6a69dd7c7/sklearn/ensemble/_forest.py#L76-L112

This PR added this new parameter to the package so that one can use it on scikit learn version >= 0.22.

arokem commented 4 years ago

LGTM! Merging.