search

deel-ai / puncc

👋 Puncc is a python library for predictive uncertainty quantification using conformal prediction.
https://deel-ai.github.io/puncc/
252 stars 14 forks source link

[Feature Request]: - EnbPI for any regression model #54

Open valeman opened 4 days ago

valeman commented 4 days ago

Module

Regression

Contact Details

No response

Feature Request

What classes of estimators does EnbPI in PUNCC works with?

The tutorial mentions RandomForest, the EnbPI model as such as published in paper is not limited to bagging estimators and it can work with any model.

Is there a gap in implementation vs the model in the paper?

If so, it would be good to have EnbPI work with any regression model classes including boosted trees (CatBoost/XGBoost/LightGBM) and scikit-learn regressors.

A minimal example

No response

Version

v0.9

Environment

- OS:
- Python version:
- Packages used version:
M-Mouhcine commented 3 days ago

Hi @valeman,

Puncc enables virtually any underlying learning algorithm and aggregation function for EnbPI, including neural networks (pytorch, tf ...), ensemble methods, ... as long as we correctly wrap them with a suitable wrapper (usually puncc.deel.api.prediction.BasePredictor). Here is an synthetic example using different models you can open in colab Open In Colab.

Let me know if I understood and answered correctly your question.

valeman commented 3 days ago

That’s great let me check it out