PUQ

"PUQ is a framework for building response surfaces and performing Uncertainty Quantification (UQ) and sensitivity analysis. It was created with the goal of making an easy to use framework that could be easily integrated and extended.

The features:

• Implemented as a Python library but can be used from the command line with a minimum of Python knowledge.
• Collects all results into a single HDF5 file.
• Implements Monte Carlo and Latin Hypercube sampling.
• For better scalability, includes a Smolyak sparse grid method.
• Builds response surfaces from sample points.
• Includes GUIs to visualize and compare PDFs and response surfaces.
• Can use PyMC to perform Bayesian calibration on input parameters.

https://github.com/c-PRIMED/puq

http://c-primed.github.io/puq/index.html