# Era Splitting - Invariant Learning for Gradient Boosted Decision Trees

**URL:** <https://forum.numer.ai/t/era-splitting-invariant-learning-for-gradient-boosted-decision-trees/6690>\
**Category:** Data Science\
**Created:** [September 27, 2023, 3:12pm UTC](https://forum.numer.ai/t/era-splitting-invariant-learning-for-gradient-boosted-decision-trees/6690 "2023-09-27T15:12:31Z")\
**Posts on this page:** 1\
**Showing post:** 4

<div class="post-metadata">

**Author:** ![perfect\_fit](https://yyz1.discourse-cdn.com/flex009/user_avatar/forum.numer.ai/perfect_fit/32/2527_2.png) [@perfect\_fit](https://forum.numer.ai/u/perfect_fit)\
**Post date:** [October 2, 2023, 4:28pm UTC](https://forum.numer.ai/t/era-splitting-invariant-learning-for-gradient-boosted-decision-trees/6690/4 "2023-10-02T16:28:07Z")

</div>

Do you have a lightweight version to install `EraHistGradientBoostingRegressor` (without having to install the full scikit-learn library with it)?

Would love to test how it integrates within scikit-learn `Pipeline`, `FeatureUnion`, `ColumnTransformer`, [NumerBlox 1.0](http://forum.numer.ai/t/preview-numerblox-1-0/6696) and other libraries compatible with scikit-learn. I know `HistGradientBoostingRegressor` works well in pipelines so the era version should also be fine, unless there are some issues because of needing to pass eras in the fit step.

We have a similar thing in [BayesianGMMTargetProcessor](https://github.com/crowdcent/numerblox/blob/rewrite/numerbloxv1/docs/targets.md#bayesiangmmtargetprocessor) where `.fit` requires an `eras` argument and got that to work with `sklearn.pipeline.Pipeline`.

---

_[View the full topic](https://forum.numer.ai/t/era-splitting-invariant-learning-for-gradient-boosted-decision-trees/6690)._