Skip to contents

Create a new likelihood-restricted prior sampler (LRPS)

Source: R/lrps-class.R
new_ernest_lrps.Rd

Nested sampling relies on generating a series of points in the prior space with increasing log-likelihood values. This is accomplished using a likelihood-restricted prior sampler (LRPS), which generates independent and identically distributed points from the prior, subject to a hard likelihood constraint.

To create your own LRPS, subclass new_ernest_lrps and provide S3 methods for propose() and update_lrps() for your subclass.

Usage

new_ernest_lrps(
  unit_log_fn = NULL,
  n_dim = NULL,
  max_loop = getOption("ernest.max_loop", 1000000L),
  cache = NULL,
  ...,
  .class = NULL,
  .call = caller_env()
)

Arguments

unit_log_fn, n_dim

Provided when ernest_sampler() is called with a given ernest_lrps:

  • unit_log_fn (function, optional): Takes a matrix of points in the unit cube and returns a numeric vector of log-likelihood values.

  • n_dim (integer, optional): Number of dimensions of the prior space.

max_loop

Positive integer. Maximum number of attempts to generate points via region-based sampling. Usually hidden from users, but can be set via the ernest.max_loop option.

cache

(environment, optional) Environment for caching values. If NULL, a new environment is created.

...

<dynamic-dots> Name-value pairs for additional elements for subclasses of this LRPS.

.class

(character vector, optional) Subclasses of this LRPS.

.call

The execution environment of a currently running function, e.g. caller_env(). The function will be mentioned in error messages as the source of the error. See the call argument of abort() for more information.

Value

An LRPS specification: a list containing the input arguments, with a class specific to the LRPS type.