pymc.smc.kernels.MH#

class pymc.smc.kernels.MH(*args, correlation_threshold=0.01, **kwargs)[source]#

Metropolis-Hastings SMC_kernel.

Methods

MH.__init__(*args[, correlation_threshold])

Create a Metropolis-Hastings SMC kernel.

MH.initialize(start, rng)

Initialize the kernel for sampling.

MH.initialize_population()

Create an initial population from the prior distribution.

MH.mutate()

Metropolis-Hastings perturbation.

MH.resample()

Resample particles based on importance weights.

MH.sample_settings()

SMC_kernel settings to be saved once at the end of sampling.

MH.sample_stats()

Stats to be saved at the end of each stage.

MH.set_rng(rng)

Copy compiled functions, updating their random number generators.

MH.setup_kernel()

Proposal dist is just a Multivariate Normal with unit identity covariance.

MH.step()

Perform a single SMC stage: resample, tune, and mutate.

MH.tune()

Update proposal scales for each particle dimension and update number of MH steps.

MH.update_beta_and_weights()

Calculate the next inverse temperature (beta).

Attributes

stats_dtypes_shapes