ssoudan / pyGP

Various experimentations around Gaussian Processes.

pyGP

Various experimentations around Gaussian Processes.

Run

conda create --name pyGP python=3.7
conda activate pyGP
pip install -r requirements.txt

and then 'run' the main's in example/ from PyCharm.

or wtih docker:

./BUILD.sh && ./RUN.sh

You'll find the output in ... output folder.

What is in there?

Don't expect too much originality here, it's mostly a bunch of tutorial or example code put together in the same place.

GP regression:

• src/main/python/run_gpflow.py: MLE and MCMC using GPflow - https://gpflow.readthedocs.io/en/stable/notebooks/regression.html
• src/main/python/run_sklearn.py: MLE using Scikit-learn - https://scikit-learn.org/stable/
• src/main/python/run_tfp.py: MLE and HMC using Tensorflow probability - https://www.tensorflow.org/probability

Bayesian optimization:

• src/main/python/run_bo.py: comparison of the different variantes for BO purpose