Learning Confidence Bounds for Classification with Imbalanced Data - ECAI 2024

Package for our paper: https://arxiv.org/abs/2407.11878

Make sure to keep up to date with our latest version of this package found at: https://github.com/mattclifford1/Confidence-Bounds-ECAI24

Setting up Environment

First clone this repo

git clone https://github.com/mattclifford1/Confidence-Bounds-ECAI24
cd Confidence-Bounds-ECAI24

Then make a python env e.g. with conda

conda create -n CliffordECAI24 python=3.10 -y
conda activate CliffordECAI24

Install as editable package

pip install -e .

Reproducing the experiments from the paper

run the file 'experiments/run_all.py'

python experiments/run_all.py 

this will save each dataset's results to experiments/results and will combine all the results into one latex table experiments/combined_table.txt

N.B. you will need to download and process and the MIMIC-ICU dataset externally as I do not have the rights to include it within this repo. Please contact me if you have any questions.

Applying to your own classifier

Check out our latest examples and documentation on how to apply to your own classifier and ways of customisation at: https://github.com/mattclifford1/Confidence-Bounds-ECAI24/blob/main/examples.md