rxmc.param_sampling.AdaptiveMetropolisSampler#

class rxmc.param_sampling.AdaptiveMetropolisSampler(params: list[Parameter], prior, starting_location: ndarray, adapt_start: int = 100, window_size: int = 1000, epsilon_fraction: float = 1e-06)[source]#

Bases: Sampler

Metropolis sampler that adapts the proposal covariance with a sliding window.

The proposal covariance is estimated from the last window_size samples after adapt_start steps have been collected.

Parameters:

  • params (list of Parameter) – Parameters to sample.

  • prior (object) – Prior distribution with a callable logpdf(x) method.

  • starting_location (np.ndarray, shape (ndim,)) – Initial parameter vector.

  • adapt_start (int, optional) – Step at which adaptation begins. Defaults to 100.

  • window_size (int, optional) – Number of past samples used for covariance estimation. Defaults to 1000.

  • epsilon_fraction (float, optional) – Small regularisation term for the proposal covariance. Defaults to 1e-6.

__init__(params: list[Parameter], prior, starting_location: ndarray, adapt_start: int = 100, window_size: int = 1000, epsilon_fraction: float = 1e-06)[source]#

Methods

__init__(params, prior, starting_location[, ...])

batch_acceptance_fractions()

Acceptance fraction for each completed batch.

most_recent_batch_acceptance_fraction()

Acceptance fraction of the most recent batch.

overall_acceptance_fraction()

Overall acceptance fraction across all completed batches.

record_batch(n_steps, n_accepted, chain, ...)

Append a completed batch to the running chain.

sample(n_steps, starting_location, rng, ...)

Run the adaptive sampler for one batch, passing history to the algorithm.
sample(n_steps: int, starting_location: ndarray, rng: Generator, log_posterior: Callable[[ndarray], float], burn: bool = False)[source]#

Run the adaptive sampler for one batch, passing history to the algorithm.

Overrides Sampler.sample() to forward the accumulated chain so the adaptive algorithm can estimate the covariance from all past samples.

Parameters:

  • n_steps (int) – Number of steps to run.

  • starting_location (np.ndarray, shape (ndim,)) – Starting parameter vector.

  • rng (np.random.Generator) – Random number generator.

  • log_posterior (callable) – Function f(x) -> float returning the log posterior.

  • burn (bool, optional) – If True, discard samples. Defaults to False.