Skip to contents

Sample from Prior Distribution (Simplified)

Source: R/sample_from_prior.R
sample_from_prior.Rd

A simplified unified function to sample values from prior distributions. Expects all priors to have consistent structure with $distribution and $parameters slots.

Usage

sample_from_prior(n = 1, prior, verbose = FALSE)

Arguments

n

Integer. Number of samples to draw (default: 1).

prior

List. Prior object containing 'distribution' and 'parameters' slots.

verbose

Logical. If TRUE, print warnings for NULL or invalid priors (default: FALSE).

Value

Numeric vector of length n containing samples from the prior distribution. Returns NA values if prior is NULL/invalid.

Details

This simplified version expects every prior to have a consistent structure:

  • $distribution: Character string naming the distribution type

  • $parameters: List containing the distribution parameters

Supported distributions:

  • beta: parameters$shape1, parameters$shape2

  • gamma: parameters$shape, parameters$rate

  • lognormal: parameters$meanlog, parameters$sdlog OR parameters$mean, parameters$sd; optional parameters$lower / parameters$upper truncate the draw (meanlog/sdlog then describe the untruncated distribution)

  • normal: parameters$mean, parameters$sd

  • truncnorm: parameters$mean, parameters$sd, parameters$a (lower bound), parameters$b (upper bound)

  • uniform: parameters$min, parameters$max

  • discrete_uniform: parameters$min, parameters$max (returns integers)

  • gompertz: parameters$b, parameters$eta

  • fixed: parameters$value (genuinely converged, returns constant)

  • frozen: parameters$value (frozen in previous stage, returns constant)

Examples

# Sample from a beta distribution
prior_beta <- list(
  distribution = "beta",
  parameters = list(shape1 = 2, shape2 = 5)
)
sample_from_prior(n = 10, prior = prior_beta)
#>  [1] 0.19007042 0.72965473 0.48529669 0.43924840 0.11377246 0.26857255
#>  [7] 0.15743964 0.15911719 0.26804349 0.08400777

# Sample from a truncated normal distribution
prior_truncnorm <- list(
  distribution = "truncnorm",
  parameters = list(mean = 0, sd = 1, a = -2, b = 2)
)
sample_from_prior(n = 10, prior = prior_truncnorm)
#>  [1] -1.08609755 -0.68370851 -1.51885919 -0.29066904  0.24961982  1.49934126
#>  [7]  0.43712317 -0.90208633  0.01721895  0.22281779