hdmrlib 1.0.0

High Dimensional Model Representation (HDMR) and Enhanced Multivariate Products Representation (EMPR) library

HDMRLib

HDMRLib is an open-source Python library for High-Dimensional Model Representation (HDMR) and Enhanced Multivariate Products Representation (EMPR). It provides a unified workflow for decomposition, component analysis, and lower-order reconstruction across NumPy, PyTorch, and TensorFlow backends.

Features

• HDMR and EMPR in one library
• Unified decomposition workflow
• Component extraction and lower-order reconstruction
• NumPy, PyTorch, and TensorFlow support
• Documentation, examples, and tests

Installation

Install the package from PyPI:

pip install hdmrlib

For development installation:

git clone https://github.com/hdmrlib/HDMRLib.git
cd HDMRLib
pip install -e .

For optional backend dependencies and development tools, see the documentation.

Quick Example

import numpy as np
from hdmrlib import HDMR

X = np.random.rand(10, 10)

model = HDMR(X, order=2)

X_reconstructed = model.reconstruct()
components = model.components()

This computes a second-order HDMR decomposition of X, reconstructs the data from the decomposition, and returns the extracted component terms.

The same workflow also applies to EMPR:

from hdmrlib import EMPR

model = EMPR(X, order=2)
X_reconstructed = model.reconstruct()
components = model.components()

Documentation

Full documentation is available at:

• Documentation website: https://hdmrlib.github.io/HDMRLib/

Supported Backends

HDMRLib currently supports:

• NumPy for standard array-based workflows
• PyTorch for tensor computation and backend-integrated workflows
• TensorFlow for TensorFlow-based numerical workflows

Backend selection is handled through a unified interface, allowing the same decomposition workflow to be used across supported numerical libraries.

Testing

To install development dependencies and run the test suite:

pip install -e ".[dev]"
python -m pytest

If you want to test optional backends as well:

pip install -e ".[dev,all]"
python -m pytest

Citation

If you use HDMRLib in academic work, please cite the associated software or publication.

@software{hdmrlib,
  title = {HDMRLib: A Python Library for HDMR and EMPR},
  author = {Pınar Yalçın Güler and Muhammed Enis Şen and Buğra Eyidoğan},
  year = {2026},
  url = {https://github.com/hdmrlib/HDMRLib}
}

If a paper citation becomes available, it should be preferred here.

Contributing

Contributions, bug reports, and feature suggestions are welcome.

Please use the GitHub issue tracker for bug reports and discussions, and open a pull request for proposed changes.

License

This project is released under the terms of the license included in the repository.