Skip to content

Latest commit

 

History

History
10 lines (6 loc) · 823 Bytes

README.md

File metadata and controls

10 lines (6 loc) · 823 Bytes

Uncertainty Aware Conformalized Quantile Regression (UACQR)

Here we provide an implementation of UACQR and notebooks to reproduce the figures and tables from "Integrating Uncertainty Awareness into Conformalized Quantile Regression"

We demonstrate how to use our method for neural networks and Quantile Regression Forests in the two demo notebooks.

For more information on the hyperparameter options for Quantile Regression Forests, see sklearn_quantile documentation: https://sklearn-quantile.readthedocs.io/en/latest/

For more information on the hyperparameter options and architectural details for our neural network implementation, see helper.py

For real data experiments, the GetDatasets module and associated assets are a fork from https://github.com/msesia/cqr-comparison