-
Hi @mnwright ,
I'm writing to request a new feature in **ranger**, if feasible. In particular, it would be extremely useful to be able to remove/trim specific trees off of the fitted `"ranger"` obj…
-
According to this article (https://christophm.github.io/interpretable-ml-book/feature-importance.html)
Permutation based features importance should be performed with test data, so am I doing right …
-
I'd suspect there's a bit of a cottage industry with tools to modify CTSM parameter files. Is this something that we should more formally support or provide tools for? This could be basic scripts in…
-
The goal is to test if Linear Regression, Logistic Regression, and Tree-based models maintain acceptable arithmetic precision.
-
## Inliner: methodology for code quality measurements
We'd like to measure the impact of inlining on performance, so we can
build predictive models. Ideally we'd measure the _isolated_ impact of
just…
-
This is very painful. I wish we would stop committing images here.
```
$ git pull
remote: Enumerating objects: 1347, done.
remote: Counting objects: 100% (1347/1347), done.
remote: Compressin…
-
In this review https://github.com/dotnet/machinelearning/pull/2753, I notice the methods to traverse thru trained tree ensembles have been removed from the public API
@wschin @TomFinley
I under…
-
The repro is [here](https://colab.research.google.com/drive/101ZT_F2_iCzRnh85Z_3_3hD4zQfJrWci?usp=sharing). In short, functions which return PyTrees involving non-JAX types break `filter_{vmap,pmap}`.…
-
Example: https://mila.quebec/en/publications/
It would be nice to reuse the same code as in the Mila website. Not sure if that's 'easily' possible via RTD
-
Currently ONNX doesn't support categorical vars which limits the use of tree ensembles in OMLT. By directly interfacing with these libraries, we can add support for categorical vars.