/ModelCalibration

Development of general interfaces to make model calibration tasks transparent

Primary LanguagePythonMIT LicenseMIT

Introduction - Model calibration

Model calibration (or model updating) is a task most scientists are facing when building parameterized models and then updating the parameters based on some experimental data in order to generalize the model and make accurate predictions. For complex models and complex data sets (different experimental tests with different data structures, sensor types and different models) this task is often tedious, difficult to reproduce and often error prone due to complex challenges related to data processing, forward model development and inference. The aim of this roject is to make the process more transparent, easier to setup and work with and more transparent when different people are jointly performing this task. Further information can be found in the documentation.

Installation of the conda environment

Clone the repository and make sure to include the submodules

git clone --recurse-submodules https://github.com/BAMresearch/ModelCalibration.git

Create environment within the main project folder

conda env create --prefix ./conda-env -f environment.yml 

Update environment within main project folder (if environment.yml was changed)

conda env update --prefix ./conda-env -f environment.yml --prune

Activate environment

conda activate ./conda-env

Install submodule Bayes

git pull --recurse-submodules
pip install -e BayesianInference