bayou (version 2.1.1)

make.powerposteriorFn: Makes a power posterior function in bayou

Description

This function generates a power posterior function for estimation of marginal likelihood using the stepping stone method

Usage

make.powerposteriorFn(Bk, priorFn, refFn, model)

Arguments

Bk

The sequence of steps to be taken from the reference function to the posterior

priorFn

The prior function to be used in marginal likelihood estimation

refFn

The reference function generated using make.refFn() from a preexisting mcmc chain

model

A string specifying the model type ("OU", "OUrepar", "QG") or a model parameter list

Value

A function of class "powerposteriorFn" that returns a list of four values: result (the log density of the power posterior), lik (the log likelihood), prior (the log prior), ref the log reference density.

Details

For use in stepping stone estimation of the marginal likelihood using the method of Fan et al. (2011).