Posterior Samples

Class providing lists of parameter-space states at which to evaluate the posterior distribution—a pre-defined grid or set of points used for posterior exploration or visualization rather than adaptive MCMC sampling. Implementations generate arrays of posteriorSampleStateSimple objects covering the parameter space according to a chosen scheme (e.g. a regular grid over the prior, or a set of previously sampled points). These samples are used to compute the posterior probability at each grid point for plotting or convergence diagnostics.

Default implementation: posteriorSamplesPriorGrid

Methods

samples

Return the array of pre-defined parameter-space states at which the posterior probability will be evaluated, allocating the simulationStates array according to the sampling scheme.

• type(posteriorSampleStateSimple)(:) simulationStates [inout]

• type(modelParameterList)(:) modelParameters_ [inout]

posteriorSamplesLatinHypercube

A posterior state samples class which draws samples from a Latin hypercube in the cumulative distribution of the priors.

Parameters

• [countSamples] (integer) — The number of samples to draw.

• [maximinTrialCount] (integer; default 1000) — The number of trial Latin Hypercubes to construct when seeking the maximum minimum separation sample.

posteriorSamplesPriorGrid

A posterior state samples class which draws samples from a grid in the cumulative distribution of the priors.

(Default implementation)

Parameters

• [countGrid] (integer) — The number of grid steps in each parameter.