get_sampler

Synopsis

Return information on the current pyBLoCXS sampler.

Syntax

get_sampler()
get_sampler_opt( [optname] )
get_sampler_name()

Description

pyBLoCXS is a sophisticated Markov chain Monte Carlo (MCMC) based algorithm designed to carry out Bayesian Low-Count X-ray Spectral (BLoCXS) analysis in the Sherpa environment. The algorithm explores parameter space at a suspected minimum - i.e. after a standard Sherpa fit.

The Sherpa get_sampler commands return information about the type of jumping rule used in MCMC.

Jumping Rules

"MH" is Metropolis-Hastings, which always jumps from the best-fit, and "MetropolisMH" is Metropolis with Metropolis-Hastings that jumps from the best-fit with probability 'p_M', else it jumps from the last accepted jump. "PragBayes" is used when effective area calibration uncertainty is to be included in the calculation. (At each nominal MCMC iteration, a new calibration product is generated, and a series of N (option in set_sampler_opt) MCMC sub-iteration steps are carried out, choosing between Metropolis and Metropolis-Hastings types of samplers with probability p_M (option in set_sampler_opt). Only the last of these sub-iterations are kept in the chain.)

get_sampler - access configuration options for the current sampler
get_sampler_opt - access a particular sampler option for a pyBLoCXS chain
get_sampler_name - access the current sampler name

The configuration options returned by get_sampler include the following:

Metropolis-Hastings Jumping Rule

defaultprior - Boolean to indicate that all parameters have the default flat prior.
inv - Boolean or array of booleans indicating which parameters are on the inverse scale.
log - Boolean or array of booleans indicating which parameters are on the logarithm scale (natural log).
originalscale - Array of booleans indicating which parameters are on the original scale.
priorshape - Array of booleans indicating which parameters have associated user-defined prior functions.
scale - A scalar multiple of the output of covar() used in the scale of the t-distribution.

Mixture of Metropolis and Metropolis-Hastings Jumping Rules

defaultprior - Boolean to indicate that all parameters have the default flat prior.
inv - Boolean or array of booleans indicating which parameters are on the inverse scale.
log - Boolean or array of booleans indicating which parameters are on the logarithm scale (natural log).
originalscale - Array of booleans indicating which parameters are on the original scale.
p_M - The proportion of jumps generated by the Metropolis jumping rule.
priorshape - Array of booleans indicating which parameters have associated user-defined prior functions.
scale - A scalar multiple of the output of covar() used in the scale of the t-distribution.

Available samplers are returned by the list_samplers command.

Refer to the pyBLoCXS documentation for additional information about the algorithm.

Examples

Example 1

sherpa> print get_sampler_name()

Return the name of the current sampler. In this example, it is the default "MetropolisMH" for Metropolis withnMetropolis-Hastings.

sherpa> load_pha("pha.fits")
sherpa> set_source(xsphabs.abs1 * powlaw1d.p1)
sherpa> set_stat("cash")
sherpa> fit()
sherpa> covar()

sherpa> print get_sampler_name()
        MetropolisMH

Example 2

sherpa> print get_sampler()

Retrieve the configuration options for the current sampler. The output for this example is:

"{'priorshape': False, 'scale': 1, 'log': False, 'defaultprior': True,
'inv': False, 'sigma_m': False, 'priors': (), 'originalscale': True,
'verbose': False}"

Example 3

sherpa> get_sampler_opt('log')

Return the current setting for the 'log' configuration option of the current sampler.

sherpa> get_sampler_opt('log')
        False

Bugs

See the bugs pages on the Sherpa website for an up-to-date listing of known bugs.