BayesValidRox

An open-source, object-oriented Python package for surrogate-assisted Bayesain Validation of computational models. This framework provides an automated workflow for surrogate-based sensitivity analysis, Bayesian calibration, and validation of computational models with a modular structure.

Features

• Surrogate modeling with Polynomial Chaos Expansion, Gaussian Process Emulator, mixed surrogate types
• Global sensitivity analysis using Sobol Indices
• Bayesian calibration and validation with Rejection sampling or MCMC using emcee package
• Bayesian model comparison with model weights or confusion matrix for multi-model setting

Resources

The following resources are useful to get started on working with BayesValidRox:

• BayesValidRox website
• User guide
• Tutorial
• Publucation

Important links:

• GitLab
• Changelog

Authors

Best to contact: @RKohlhaas, @mariafer.morales

Full list of authors:

• @farid
• @RKohlhaas
• @mariafer.morales
• @alacheim

Installation

The best practive is to create a virtual environment and install the package inside it.

To create and activate the virtual environment run the following command in the terminal:

  python3 -m venv bayes_env
  cd bayes_env
  source bin/activate

You can replace bayes_env with your preferred name. For more information on virtual environments see this link.

Now, you can install the latest release of the package on PyPI inside the venv with:

  pip install bayesvalidrox

and installing the version on the master branch can be done by cloning this repo and installing:

  git clone https://git.iws.uni-stuttgart.de/inversemodeling/bayesvalidrox.git
  cd bayesvalidrox
  pip install .

Requirements

Core packages (same for all Python >=3.9)

• numpy>=1.26.4
• pandas>=2.2.3
• joblib>=1.5.3
• seaborn>=0.13.2
• tqdm>=4.64.0
• chaospy==4.3.21
• h5py>=3.9.0
• statsmodels>=0.14.2
• multiprocess>=0.70.16
• corner==2.2.3

Python 3.9

• matplotlib>=3.7,<3.10
• scipy>=1.9,<1.11
• emcee>=3.0.2,<3.1.6
• scikit-learn>=1.0,<1.2
• datasets>=2.20.0,<4.0.0

Python >= 3.10

• matplotlib>=3.10.0
• scipy>=1.15.2
• emcee>=3.1.6
• scikit-learn>=1.7.2,<1.8.0
• datasets>=4.6.1

TexLive for Plotting with matplotlib

Here you need super user rights

sudo apt-get install dvipng texlive-latex-extra texlive-fonts-recommended cm-super