UniBM

UniBM is a Python toolkit for extreme risk modeling and statistical analysis, with a focus on methods like sub-sampling block maxima as exampled in vignette.ipynb. The functions in this repository are designed to accompany the methods detailed in our research paper. For in-depth explanations and applications, users are strongly encouraged to consult the paper.

All bug reports and feature requests are welcomed.

The scripts are licensed under the MIT License.

Citation

If you use unibm in your work, please cite:

Cheng, T., Peng, X., Choiruddin, A., He, X., & Chen, K. (2025). Environmental extreme risk modeling via sub-sampling block maxima. arXiv preprint arXiv:2506.14556.

@article{cheng2025environmental,
title={Environmental extreme risk modeling via sub-sampling block maxima},
author={Cheng, Tuoyuan and Peng, Xiao and Choiruddin, Achmad and He, Xiaogang and Chen, Kan},
journal={arXiv preprint arXiv:2506.14556},
year={2025}
}

(Recommended) uv for Dependency Management and Packaging

After git clone https://github.com/TY-Cheng/UniBM.git, cd into the project root where pyproject.toml exists,

# From inside the project root folder
# Sync dependencies with CPU support (default)
uv sync --extra cpu

Functions

unibm

cdf_func_kernel non parametric cdf estimator, by kernel smoothing

est_tail_dep_coeff pairwise tail dependence coefficient estimator

est_extremal_index_reciprocal extremal index (EI) estimator, by reciprocal of the mean of the block maxima;

viz_eir chart results from est_extremal_index_reciprocal(is_retn_vec=True)

est_extreme_value_index extreme value index (EVI) estimator, by MPMR & EMR

viz_evi_reg chart results from est_extreme_value_index(is_retn_vec=True)

License

This project is released under the MIT License (© 2024- Tuoyuan Cheng, Kan Chen).