Fit Bayesian generalised (non-)linear multilevel compositional model via full Bayesian inference

Fit a brm model with multilevel ILR coordinates

Usage

brmcoda(complr, formula, ...)

Arguments

complr: A complr object containing data of composition, ILR coordinates, and other variables used in the model.
formula: A object of class formula, brmsformula: A symbolic description of the model to be fitted. Details of the model specification can be found in brmsformula.
...: Further arguments passed to brm.

Value

A brmcoda with two elements

complr: An object of class complr used in the brm model.
model: An object of class brmsfit, which contains the posterior draws along with many other useful information about the model.

Examples

# \donttest{
if(requireNamespace("cmdstanr")){
  cilr <- complr(data = mcompd, sbp = sbp,
                 parts = c("TST", "WAKE", "MVPA", "LPA", "SB"), idvar = "ID")
  
  # inspects ILRs before passing to brmcoda
  names(cilr$between_logratio)
  names(cilr$within_logratio)
  names(cilr$logratio)
  
  # model with compositional predictor at between and within-person levels
  m1 <- brmcoda(complr = cilr,
                formula = Stress ~ bilr1 + bilr2 + bilr3 + bilr4 +
                                   wilr1 + wilr2 + wilr3 + wilr4 + (1 | ID),
                chain = 1, iter = 500,
                backend = "cmdstanr")
  
  # model with compositional outcome
  m2 <- brmcoda(complr = cilr,
                formula = mvbind(ilr1, ilr2, ilr3, ilr4) ~ Stress + Female + (1 | ID),
                chain = 1, iter = 500,
                backend = "cmdstanr")
  }# }
#> Error: CmdStan path has not been set yet. See ?set_cmdstan_path.