set_MH: Set options for Metropolis-Hastings sampling

In mcmcsae: Markov Chain Monte Carlo Small Area Estimation

Set options for Metropolis-Hastings sampling

Description

Set options for Metropolis-Hastings sampling

Usage

set_MH(type = "RWTN", scale = 0.025, adaptive = NULL, ...)

Arguments

type	a character string defining the proposal distribution. Among supported types are random walk proposals "RWTN", "RWN" and "RWLN" with truncated normal, normal and log-normal proposal distributions. Other choices correspond to independence proposals: "TN" for a truncated normal proposal, "unif" for a uniform proposal, and "beta" and "gamma" for specific beta and gamma proposal distributions. Not all types are supported for a particular parameter; see the specific help of the function defining the model component of interest to see which proposal distribution types are supported.
scale	in case of the "RWTN" proposal, the (initial) scale of the distribution.
adaptive	in case of the random walk "RWTN" or "RWN" proposals, whether the scale parameter is adapted based on acceptance rates during the burnin phase of the MCMC simulation. The default is TRUE in these cases.
...	additional parameters depending on the proposal type. Supported arguments are 'l' and 'u' to pass the lower and upper limits of uniform or random walk truncated normal proposals (defaults l=0 and u=1), and 'a' and 'b' to pass the shape parameters of a beta proposal distribution (defaults a = b = 0.5).

Value

An environment with variables and methods for Metropolis-Hastings sampling, for use by other package functions.

mcmcsae documentation built on April 12, 2025, 2:25 a.m.