.. _method-bayes_calibration-sub_sampling_period:

"""""""""""""""""""
sub_sampling_period
"""""""""""""""""""


Specify a sub-sampling of the MCMC chain


.. toctree::
   :hidden:
   :maxdepth: 1



**Specification**

- *Alias:* None

- *Arguments:* INTEGER

- *Default:* 0


**Description**


If a ``sub_sampling_period`` is specified, the MCMC chain is filtered such that
only the sample at the beginning of each period is in the final MCMC chain.
The ``sub_sampling_period`` should therefore be greater than or equal to the
correlation length of the samples.

*Default Behavior*

If not specified, all MCMC samples are included in the final chain.

*Expected Output*

If specified, an additional tabular output file containing the final (sub-
sampled) MCMC chain is generated.



**Examples**


Generate and then filter a 1000 sample MCMC chain, retainining the
first accepted sample and every 100th sample thereafter.

An example method block of a Dakota input file is given below


.. code-block::

    method,
     bayes_calibration queso
       chain_samples = 1000 seed = 348
       dram
     proposal_covariance
       diagonal values 1.0e6 1.0e-1
     sub_sampling_period 100