elicits 0.0.7

Tool for learning prior distributions based on expert knowledge

Learning prior distributions based on expert knowledge (Expert knowledge elicitation)

Note: This project is still in the development stage and not yet tested for practical use.

Description

The elicit package provides a simulation-based framework for learning either parametric or non-parametric, as well as independent or join prior distributions for parameters in a Bayesian model based on expert knowledge.

Further information can be found in the corresponding papers:

• Bockting, F., Radev S. T., & Bürkner P. C. (2024) Expert-elicitation method for non-parametric joint priors using normalizing flows. Preprint at https://arxiv.org/abs/2411.15826
• Bockting, F., Radev, S. T. & Bürkner, P. C. (2024). Simulation-based prior knowledge elicitation for parametric Bayesian models. Scientific Reports 14, 17330 (2024). https://doi.org/10.1038/s41598-024-68090-7

Installation

• requires: Python >=3.10 and < 3.12

Via pip and virtual environments

If you want to use a python environment (here with virtualenv)

# create an environment
virtualenv elicit-env python=python3.11

# activate it
source elicit-env/Scripts/activate 

Another option is the use of a conda environment

# create an environment
conda create --name=elicit-env python==3.11

# activate it
conda activate elicit-env

Install package via pip

# install elicit package
pip install elicits

Install from GitHub

Install elicit from GitHub via

pip install git+https://github.com/florence-bockting/elicit

If you need access to the source code, instead use

git clone [e-mail]:florence-bockting/elicit.git
cd elicit
pip install -e .

Usage

See our project website with tutorials for usage examples.

License

This work is licensed under multiple licences:

• All original source code is licensed under Apache License 2.0.
• All documentation is licensed under CC-BY-SA-4.0.

Documentation

Documentation for this project can be found on the project website.

Citation and Reference

This work builds on the following references

• Bockting, F., Radev, S. T., & Bürkner, P. C. (2024). Simulation-based prior knowledge elicitation for parametric Bayesian models. Scientific Reports, 14(1), 17330. (see PDF)
• Bockting, F., Radev S. T., & Bürkner P. C. (2024) Expert-elicitation method for non-parametric joint priors using normalizing flows. Preprint at https://arxiv.org/abs/2411.15826

BibTeX:

@article{bockting2024simulation,
  title={Simulation-based prior knowledge elicitation for parametric Bayesian models},
  author={Bockting, Florence and Radev, Stefan T and B{\"u}rkner, Paul-Christian},
  journal={Scientific Reports},
  volume={14},
  number={1},
  pages={17330},
  year={2024},
  doi={10.1038/s41598-024-68090-7},
  publisher={Nature Publishing Group UK London}
}

@article{bockting2024expert,
  title={Expert-elicitation method for non-parametric joint priors using normalizing flows},
  author={Bockting, Florence and Radev, Stefan T and B{\"u}rkner, Paul-Christian},
  journal={arXiv preprint},
  year={2024},
  doi={https://arxiv.org/abs/2411.15826}
}

Authors and Contributors

You are very welcome to contribute to our project. If you find an issue or have a feature request, please use our issue templates. For those of you who would like to contribute to our project, please have a look at our contributing guidelines.

Authors

Florence Bockting

Paul-Christian Bürkner

Contributors

Luna Fazio 🖋

Stefan T. Radev 🖋

This project follows the all-contributors specification. Contributions of any kind welcome!