neilzxu / robust_weighted_classification

Code for "Class-Weighted Classification: Trade-offs and Robust Approaches"

Home Page:https://arxiv.org/abs/2005.12914

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

Code for Class-Weighted Classification: Trade-offs and Robust Approaches

This repo provides supporting Python code for the paper

Xu, Z., Dan, C., Khim, J., Ravikumar, P. (2020). Class-Weighted Classification: Trade-offs and Robust Approaches. arXiv preprint arXiv:2005.12914.

Setup

Requires conda with Python 3.7.

  1. Install conda dependencies in the environment: conda env create -f environment.yml
  2. Run download_uci_data.sh from the repo main directory to download the Covertype dataset.
  3. Activate the conda enviroment with conda activate robust_weighting
  4. Setup up a wandb account and create a project named extreme-classification (or rename the project argument inside the wandb.init call inside src/main.py )

Scripts

Navigate to the root of the repo.

Run ./power_exp.sh for the synthetic experiment results.

Run ./uci_exp.sh for the real world dataset (Covertype) results.

View results in the wandb website.

About

Code for "Class-Weighted Classification: Trade-offs and Robust Approaches"

https://arxiv.org/abs/2005.12914


Languages

Language:Python 90.8%Language:Shell 9.2%