wellwise: WELL Water, Infants, and a Safe Environment: simulations

Description Usage Arguments Details Examples

Fit a Stan/JAGS model in a single data set

data_analyst(i, raw = raw, fl = "~/temp.csv", s.code = NULL,
  s.file = NULL, iter = 2500, warmup = 500, chains = 4, p = NULL,
  dx = NULL, N = NULL, env = NULL, verbose = FALSE,
  debug = FALSE, type = c("stan", "jags", "julia"), stanfit = NA,
  basis = c("identity", "iqr", "sd"), juliabin = NULL, ...)

`i`	iteration in simulated data
`raw`	data frame with raw data
`fl`	Filename to output model results for single run (useful for debugging)
`s.code`	name of a character string with a stan model
`s.file`	(experimental) character name of a presepecified stan model ("wwbd_simtemplate_20190205", "wwbd_simlogistic_TransParameter_0211201920190211")
`iter`	Number of "sweep" iterations, or iterations after the burnin/warmup
`warmup`	warmup (stan) / n.adapt (jags). Number of iterations to allow for adaptation of MCMC parameters/burnin
`chains`	Number of parallel MCMC chains (default 4)
`p`	number of beta parameters in outcome model (optional - R makes a guess)
`dx`	number of columns in data matrix (optional - R makes a guess)
`N`	Sample size (optional - R makes a guess)
`env`	environment for existing julia instance
`verbose`	print extra debugging info? FALSE
`debug`	print extra debugging info? FALSE
`type`	Which type of model is it? Can be 'stan' or 'jags'
`stanfit`	Name of stan_model or stan output that can be recycled. This will use a pre-compiled version of the stan code which cuts simulation time significantly over multiple runs.
`basis`	how to treat exposure variables: "identity" = do not transform (default); "iqr"= divide every exposure variable by its interquartile range "sd" = divide every exposure variable by its sd Note that the parameterization of the BGF interventions will be invariant to these transformations, so they change the strength of the priors, but no modifications are necessary to examine interventions of the type: "reduce exposure j by p of exposure require that the basis be "identity"
`juliabin`	NULL or path to directory containing the julia binary file e.g. /Applications/Julia-1.1.app/Contents/Resources/julia/bin/ on mac C:/Users/<username>]/AppData/Local/Julia-1.1.0/bin/ on windows 10
`...`	arguments to stan(), coda.samples(), or julia.sample() for a Stan, JAGS, or julia model, respectively