This repository contains the code necessary to run the calculations presented in our paper, Fast and Interpretable Machine Learning Modelling of Atmospheric Molecular Clusters.
Preprint of the ExplainReduce paper
Seppäläinen, Lauri, Kube\v{c}ka, Jakub, Elm, Jonas, and Puolamäki, Kai (2025).
Fast and Interpretable Machine Learning Modelling of Atmospheric Molecular Clusters
Arxiv preprint https://doi.org/10.48550/arXiv.2509.11728.
To run the calculations, a working and up-to-date installation of JKCS (version 2.1 or higher) is needed.
Add the path to your installation to config.sh.
For other dependencies (for plotting etc.), we recommend the uv package manager. You can install the dependencies by running
uv sync
after cloning the repository. Alternatively, install the requirements via pip with
pip3 install -r requirements.txt
The datasets used are available at ACDB.
To recreate the calculations from scratch, add your local path for ACDB to config.sh.
This repository contains the necessary files to
- recreate the calculations from scratch (i.e., the ACDB datasets)
- produce the final plots as they appear in the paper from aggregated results Intermediary result files are omitted from the repository due to their size.
Detailed instructions for running the calculations can be found in the experiments directory.