Conformal Clustering

This repository contains the code for the conformal clustering method and the reproducible experiments used in the accompanying paper.

Repository Overview

• environment.yml defines the Conda environment used for reproducibility. It pins the Python version and package dependencies needed to run the code and notebooks in this repository.
• conformal_clustering/ contains the Python package that implements our method.
• Reproducible_Code/ contains the Jupyter notebooks used to reproduce the figures in the paper.

Setup

Create the Conda environment with:

conda env create -f environment.yml
conda activate conformal_clustering

This environment is intended to provide a reproducible software setup for rerunning the experiments and figures.

Reproducing the Paper Figures

The notebooks in Reproducible_Code/ reproduce the figures and experiments from the paper. Before running them, download the data from https://zenodo.org/records/19410650 and save it in Reproducible_Code/Data/.

The Reproducible_Code/Figures/ and Reproducible_Code/Results/ directories contain generated outputs associated with these experiments.

Using the Package

The core implementation lives in conformal_clustering/. You can use this package as the main codebase for running the conformal clustering method and related utilities in your own scripts or notebooks within this repository.

Contents

• conformal_clustering/: implementation of the conformal clustering method
• Reproducible_Code/: notebooks for reproducing paper figures
• environment.yml: reproducible Conda environment specification