jexplore.steps.rwalk
====================

.. py:module:: jexplore.steps.rwalk

.. autoapi-nested-parse::

   This module contains definitions of Metropolis-Hasting all chains
   steps with random walks proposals.



Classes
-------

.. autoapisummary::

   jexplore.steps.rwalk.MVRandomWalk
   jexplore.steps.rwalk.GaussianRandomWalk
   jexplore.steps.rwalk.StudentTRandomWalk


Module Contents
---------------

.. py:class:: MVRandomWalk(dist, mask = None, scale = 1.19, **opts)

   Bases: :py:obj:`jexplore.steps.direct.Direct`


   Generic multivariate random walk proposal.

   :param dist: distribution class
   :param mask: proposal dimensions mask (default all space)
   :param scale: random walk scaling factor.
   :param \**opts: options to be passed to the distribution creator.


   .. py:attribute:: lower
      :type:  jax.Array

      covariance cholesky decomposition matrices `(nchains, dim, dim)`.


   .. py:attribute:: cd_const
      :type:  jax.Array

      effective scaling factor. This is the `scale` attribute
      divided by the square root of the space dimensionality.


   .. py:attribute:: scale
      :type:  float

      random walk scaling factor.


   .. py:method:: build(epoch)

      Step epoch initialisation method. This extends
      method :py:attr:`jexplore.steps.step.Step.build` by defining
      the `lower` and `cd_const` attributes.

      `epoch` and `sampling` attributes.
      :param epoch: current epoch.



   .. py:method:: proposal(key, state)

      Samples from :py:attr:`jexplore.steps.rwalk.MVRandomWalk.epoch_dist`
      are used to perform random walk steps on all chains.

      :param key: PRNG key used as the random key.
      :param state: a (nwalkers * ntemps, dim) array representing
                    the state point of the MCMC sampler. This corresponds
                    to the :py:attr:`jexplore.sampling.state.State.p`
                    attribute of a :py:attr:`jexplore.sampling.state.State`
                    object.

      :return: the updated PRNG key, the new state point (nwalkers * ntemps, dim)
               and a (nwalkers *ntemps) array with the log of the ratio between the
               backward and the forward transition probabilities.



.. py:class:: GaussianRandomWalk(mask = None)

   Bases: :py:obj:`MVRandomWalk`


   Gaussian random walk proposal.

   :param mask: proposal dimensions mask (default all space)


.. py:class:: StudentTRandomWalk(mask = None, nu=5.0)

   Bases: :py:obj:`MVRandomWalk`


   Student-T random walk proposal.

   :param mask: proposal dimensions mask (default all space)
   :param float nu: Student-T nu parameter (default: 5)