kmbayes_continue: Continue sampling from existing bkmr fit

In bkmrhat: Parallel Chain Tools for Bayesian Kernel Machine Regression

Continue sampling from existing bkmr fit

Description

Use this when you've used MCMC sampling with the kmbayes function, but you did not take enough samples and do not want to start over.

Usage

kmbayes_continue(fit, ...)

Arguments

fit	output from kmbayes
...	arguments to kmbayes_continue

Details

Note this does not fully start from the prior values of the MCMC chains. The kmbayes function does not allow full specification of the kernel function parameters, so this will restart the chain at the last values of all fixed effect parameters, and start the kernel r parmeters at the arithmetic mean of all r parameters from the last step in the previous chain.

Value

a bkmrfit.continued object, which inherits from bkmrfit objects similar to kmbayes output, and which can be used to make inference using functions from the bkmr package.

See Also

kmbayes_parallel

Examples

set.seed(111)
dat <- bkmr::SimData(n = 50, M = 4)
y <- dat$y
Z <- dat$Z
X <- dat$X
## Not run: 
fitty1 = bkmr::kmbayes(y=y,Z=Z,X=X, est.h=TRUE, iter=100)
# do some diagnostics here to see if 100 iterations (default) is enough
# add 100 additional iterations (for illustration - still will not be enough)
fitty2 = kmbayes_continue(fitty1, iter=100)
cobj = as.mcmc(fitty2)
varnames(cobj)


## End(Not run)

bkmrhat documentation built on March 29, 2022, 9:08 a.m.