Welcome to hetGPy’s documentation!
This landing page contains links to a set of example notebooks and the API reference.
Statement of Need
hetGPy is a Python package for heteroskedastic Gaussian Process modeling. It is a Python port of the hetGP R package.
hetGPy can be used for Gaussian Process regression, surrogate (emulator) modeling, and Bayesian Optimization.
The primary audience is anyone who uses Gaussian Process modeling, but is especially designed for modeling computer experiments where models exhibit heteroskedasticity (i.e. having a non-constant noise structure).
Python is a popular language for computer experiments and simulation. Thus, we hope hetGPy will be a valuable addition to the Python and simulation communities.
For questions, please contact: David O’Gara Division of Computational and Data Sciences, Washington University in St. Louis david.ogara@wustl.edu
Installation
hetGPy is available via PyPI (pip install hetgpy)
Or from Github.
Examples:
Reference: