SKBEL

Bayesian Evidential Learning framework built on top of scikit-learn

SKBEL: Bayesian Evidential Learning Framework

SKBEL is a Python framework for Bayesian Evidential Learning (BEL) built on top of scikit-learn. It provides tools for experimental design and uncertainty quantification in geosciences and engineering applications.

Key Features

  • Bayesian Experimental Design: Minimize posterior uncertainty by optimally selecting measurement locations
  • Integration with scikit-learn: Leverage the extensive ML ecosystem while maintaining scientific rigor
  • Geoscience Applications: Specifically designed for subsurface characterization and monitoring
  • Open Source: Fully documented and maintained on GitHub

Applications

The framework has been successfully applied to:

  • Wellhead Protection Area Design: Optimizing monitoring well placement for groundwater protection
  • Temperature Monitoring: Comparing wells vs. geophysical data for 4D temperature field monitoring
  • Subsurface Characterization: General framework for any inverse problem requiring uncertainty quantification

Technical Details

SKBEL implements the theoretical framework developed during my PhD research, providing:

  • Efficient posterior sampling using surrogate models
  • Information-theoretic experimental design criteria
  • Integration with various forward models
  • Visualization and analysis tools

Publications

This work has resulted in peer-reviewed publications in:

  • Journal of Hydrology (2021): Framework development and wellhead protection applications
  • Water Resources Research (2022): Comparison of well and geophysical data for temperature monitoring

Repository

GitHub: github.com/robinthibaut/skbel
DOI: 10.5281/zenodo.6205242
Documentation: Available in the repository

References