drake_mcmc: drake_mcmc

In nulloa/asgForecast: Importance Sampling with Assymetrical Gaussian Distribution

Description

A wrapper function to make running mcmc fns on drake easier

Usage

drake_mcmc(d, iter = 4000, warmup = iter/2, chains = 3, thin = 1,
  whichmodel = NULL)

Arguments

iter	number of interations to be run (default=2000)
chains	number of chains to be run (default=3)
thin	when you want to thin (default=10)
whichmodel	characeter of which model you want to run
dat	response variable which follows binomial dist
clusters	number of clusters to be used (default=nchains)
burnin	number of samples to be used as burnin (technically adaption, see link below)
inits	expects TRUE/FALSE, if TRUE will use maximum likelihood to get starting values. Needs 3 chains.

Value

A MCMC object

See Also

http://www.mikemeredith.net/blog/2016/Adapt_or_burn.htm

Examples

drake_mcmc(d, iter = 1000, chains = 3, clusters=3, warmup = iter/2, thin=1, whichmodel=="asg_common")

nulloa/asgForecast documentation built on July 26, 2020, 7:10 p.m.