jesterTOV.inference.samplers module

MCMC and nested sampling algorithms for Bayesian inference.

Submodules

The samplers taken from blackjax are implemented in a separate submodule. This contains the sequential Monte Carlo sampler, with Gaussian random walk MCMC kernel and NUTS MCMC kernel (note: NUTS is experimental), and the blackjax nested sampler with acceptance walk method.

Detailed documentations can be found in the following pages:

• jesterTOV.inference.samplers.blackjax module

Sampler Classes

These refer to the base class (JesterSampler) and the output class (SamplerOutput) for all samplers implemented in jesterTOV.inference.samplers. Moreover, the flowMC sampler is documented here as well.

jester_sampler.JesterSampler(likelihood, prior)	Lightweight base class for JESTER samplers.
jester_sampler.SamplerOutput(samples, log_prob)	Standardized output from JESTER samplers.
flowmc.FlowMCSampler(likelihood, prior, config)	FlowMC-specific sampler implementation.