SMCABC

class SMCABC(simulator, prior, distance='l2', requires_iid_data=None, distance_kwargs=None, num_workers=1, simulation_batch_size=1, distance_batch_size=-1, show_progress_bars=True, kernel='gaussian', algorithm_variant='C')

Bases: ABCBASE

Sequential Monte Carlo Approximate Bayesian Computation.

Parameters:

• simulator (Callable)

• prior (Distribution)

• distance (str | Callable)

• requires_iid_data (bool | None)

• distance_kwargs (Dict | None)

• num_workers (int)

• simulation_batch_size (int)

• distance_batch_size (int)

• show_progress_bars (bool)

• kernel (str | None)

• algorithm_variant (str)

static sample_from_population_with_weights(particles, weights, num_samples=1)

Return samples from particles sampled with weights.

Parameters:

• particles (Tensor)

• weights (Tensor)

• num_samples (int)

Return type:

Tensor