Skip to contents

Priors for a Bayesian quantile mixed model

Source: R/priors.R
bqmm_prior.Rd

Builds the list of prior hyperparameters passed to Stan. Defaults are weakly informative and scaled to the data (see bqmm_default_priors()); any element supplied here overrides the default.

Usage

bqmm_prior(
  beta_mean = 0,
  beta_sd = NULL,
  sigma_scale = NULL,
  re_scale = NULL,
  lkj = 2
)

Arguments

beta_mean

Numeric scalar or vector: prior mean(s) for the fixed-effect coefficients. Recycled to the number of columns of the design matrix. Default 0.

beta_sd

Positive scalar or vector: prior SD(s) for the fixed-effect coefficients. NULL (default) uses a data-scaled value.

sigma_scale

Positive scalar: half-normal scale for the ALD scale sigma. NULL (default) uses a data-scaled value.

re_scale

Positive scalar: half-normal scale for the random-effect standard deviations. NULL (default) uses a data-scaled value.

lkj

Positive scalar: LKJ shape parameter for the random-effect correlation matrix (used only when cov = "unstructured"). 2 favours weak correlations; 1 is uniform over correlation matrices.

Value

An object of class "bqmm_prior".

Examples

bqmm_prior(beta_sd = 5)
#> <bqmm_prior>
#>   beta_mean    0
#>   beta_sd      5
#>   sigma_scale  data-scaled default
#>   re_scale     data-scaled default
#>   lkj          2