Adding bayesian optimization, new tests

dflemin3 / approxposterior

A Python package for approximate Bayesian inference and optimization using Gaussian processes

https://dflemin3.github.io/approxposterior/

MIT License

41 stars 9 forks source link

Closed dflemin3 closed 4 years ago

dflemin3 commented 4 years ago

This PR adds a host of bug fixes and improvements including the following:

Bayesian optimization (bayesopt) method based on the "Expected Improvement" paradigm
1D and 2D tests for bayesopt
Routine to estimate the Monte Carlo error, test for this function
Numerous documentation improvements
New test to ensure test functions in likelihood.py don't break