egf_prior function

Prior Distributions

Functions used by egf to specify prior distributions of bottom level mixed effects model parameters.

Normal(mu = 0, sigma = 1)
LKJ(eta = 1)
Wishart(df, scale, tol = 1e-06)
InverseWishart(df, scale, tol = 1e-06)

Arguments

• mu: a numeric vector listing means.
• sigma: a positive numeric vector listing standard deviations.
• eta: a positive numeric vector listing values for the shape parameter, with 1 corresponding to a uniform distribution over the space of real, symmetric, positive definite matrices with unit diagonal elements. Lesser (greater) values concentrate the probability density around such matrices whose determinant is nearer to 0 (1).
• df: a numeric vector listing degrees of freedom. df must be greater than nrow(scale) - 1. If either df or scale has length greater than 1, then this condition is checked elementwise after recycling.
• scale: a list of real, symmetric, positive definite matrices or a matrix to be placed in a list of length 1.
• tol: a non-negative number specifying a tolerance for indefiniteness of scale. All eigenvalues of scale must exceed -tol * rho, where rho is the spectral radius of scale. However, regardless of tol, diag(scale) must be positive, as standard deviations are stored on a logarithmic scale.

Returns

A list inheriting from class egf_prior, with elements:

• family: a character string specifying a family of distributions.

• parameters: a named list of numeric vectors specifying parameter values.

Examples

Normal(mu = 0, sigma = 1)
Normal(mu = -5:5, sigma = c(0.1, 1))

LKJ(eta = 2)

u <- matrix(rnorm(9L), 3L, 3L)
utu <- crossprod(u)
uut <- tcrossprod(u)
Wishart(df = 6, scale = utu)
InverseWishart(df = 6, scale = list(utu, uut))