PC-prior for dof in a standarized Student-t

A function to evaluate the PC-prior for the degrees of freedom in a standarized Student-t distribution

inla.pc.ddof(dof, lambda, u, alpha, log=FALSE)

Arguments

dof

Degrees of freedom

log

Logical. Return the density or the log-density

lambda

The optional value of lambda, instead of defining it implicitely through u and alpha

u

The upper value of dof used to elicitate lambda, Prob(dof < u) = alpha

alpha

The probability alpha used to elicitate lambda

Details

These functions implements the PC-prior for the dof in a standarized Student-t distribution (ie. with unit variance and dof > 2). Either lambda, or u AND alpha must be given. Due the internal tabulation, dof must be larger than 2.0025.

Value

inla.pc.ddof returns the prior density for given dof.

Author

Havard Rue hrue@r-inla.org

Examples