infer (pmrf.infer)

Bayesian inference of RF models.

Provides samplers and routines to compute the posterior joint distributions of model parameters.

Classes

AbstractJointSampler()	Interface for samplers exploring the joint log-posterior (e.g. MCMC-based NUTS or HMC).
AbstractSplitSampler()	Interface for samplers needing separate likelihood and prior densities (e.g., modern Nested Sampling).
AbstractHypercubeSampler()	Interface for samplers operating in a unit hypercube (e.g., classical Nested Sampling).
InferResult(best_model, best_loglikelihood)	The result of an inference run.
SampleResult(samples, fn_values[, weights, ...])	Lower-level solver result returning from a sampling run.
PolyChord(nlive, num_repeats, nprior, nfail, ...)	The PolyChord Nested Sampler wrapped in a JAX interface.
HMC([num_warmup, target_acceptance_rate, ...])	Hamiltonian Monte Carlo (HMC) in JAX.
NUTS([num_warmup, target_acceptance_rate])	No-U-Turn Sampler (NUTS) in JAX.
NSS([num_delete, num_inner_steps, ...])	(experimental) A Nested Slice Sampler (NSS) in JAX.

Functions

is_sampler(x)	Returns if a solver is suitable for Bayesian sampling in pmrf.infer.sample.
is_inferer(x)	Returns if a solver is suitable for Bayesian inference in pmrf.infer.
sample(loglikelihood, model, frequency, ...)	Samples a given log likelihood function for a model over a frequency range.