Generator Functions for Priors in MCMCglmm

Functions for generating (co)variance matrix prior specifications in MCMCglmm that result in specified inverse-Wishart, inverse-gamma or central-\(F\) marginal priors for the variances.

Usage

IW(V=1, nu=0.002)

  IG(shape=0.001, scale=0.001)
  
  F(df2=1, scale=1000)
  
  tSD(df=1, scale=sqrt(1000))

Arguments

V: expected varaince as nu tends to infinity in a scalar inverse-Wishart prior
nu: degrees of freedom in a scalar inverse-Wishart prior
shape: shape parameter of the inverse-gamma prior
scale: scale parameter of the inverse-gamma, the scaled-F or scaled half-t
df: degrees of freedom for the half-t prior on the standard deviation
df2: denominator degrees of freedom for F prior (numerator degree-of-freedom is one)

Details

Each genertor function returns a function that generates a list of prior arguments need to specific (co)variance matrix priors in MCMCglmm. Those prior arguments result in the marginal distributions for the variancess being those specified in generator function. Since the appropriate prior arguments depend on the dimension of the (co)variance matrix, they are evalauted at run time once the dimension is determined using resove_prior.

Value

function of class prior_generator

Author

Jarrod Hadfield j.hadfield@ed.ac.uk

Examples