stan_inits: Calculate Initial Values for Stan VM Model
In ElkhartGroup/varian: Variability Analysis in R

View source: R/modeler.R

stan_inits

R Documentation

Calculate Initial Values for Stan VM Model

Description

Internal function used to get rough starting values for a variability model in Stan. Uses inidivudal standard deviations, means, and linear regressions.

Usage

stan_inits(
  stan.data,
  design = c("V -> Y", "V -> M -> Y", "V", "X -> V", "X -> V -> Y", "X -> M -> V"),
  useU = TRUE,
  UQ = FALSE,
  IIVQ = FALSE,
  centerU = 0,
  centerIIV = 0,
  ...
)

Arguments

`stan.data`	A list containing the data to be passed to Stan
`design`	A character string indicating the type of model to be run. One of “V -> Y” for variability predicting an outcome, “V -> M -> Y” for mediation of variability on an outcome, “V” to take posterior samples of individual variability estimates alone.
`useU`	A logical value whether the latent intercept estimated in Stage 1 should also be used as a predictor. Defaults to `TRUE`. Note if there is a mediator as well as main outcome, the latent intercepts will be used as a predictor for both.
`UQ`	A logical value whether the latent intercept estimated in Stage 1 should also be used as a predictor with a qudratice effect. Defaults to `FALSE`. Note if there is a mediator as well as main outcome, the latent intercepts will be used as a predictor for both.
`IIVQ`	A logical value whether the latent variabilities estimated in Stage 1 should also be used as a predictor with a qudratice effect. Defaults to `FALSE`. Note if there is a mediator as well as main outcome, the latent intercepts will be used as a predictor for both.
`centerU`	A numeric vector of length one (scalar) that is used to center the latent intercept estimates before using as a predictor of the outcome / mediator. Uses the formula: (U - centerU). Particularly useful when including quadratic terms.
`centerIIV`	A numeric vector of length one (scalar) that is used to center the latent variability estimates before using as a predictor of the outcome / mediator. Uses the formula: (IIV - centerIIV). Particularly useful when including quadratic terms.
`...`	Additional arguments passed to `sampling`.