metropolis.control {metropolis} | R Documentation |
metropolis.control
Description
metropolis.control
Usage
metropolis.control(
adapt.start = 25,
adapt.window = 200,
adapt.update = 25,
min.sigma = 0.001,
prop.sigma.start = 1,
scale = 2.4
)
Arguments
adapt.start |
start adapting after this many iterations; set to iter+1 to turn off adaptation |
adapt.window |
base acceptance rate on maximum of this many iterations |
adapt.update |
frequency of adaptation |
min.sigma |
minimum of the proposal distribution standard deviation (if set to zero, posterior may get stuck) |
prop.sigma.start |
starting value, or fixed value for proposal distribution s standard deviation |
scale |
scale value for adaptation (how much should the posterior variance estimate be scaled by?). Scale/sqrt(p) is used in metropolis_glm function, and Gelman et al. (2014, ISBN: 9781584883883) recommend a scale of 2.4 @return A list of parameters used in fitting with the following named objects adapt.start, adapt.window,adapt.update,min.sigma,prop.sigma.start,scale |
[Package metropolis version 0.1.8 Index]