geoR (version 1.5-6)

sample.prior: Sample the prior distribution

Description

Sample quadruples $(\beta, \sigma^2, \phi, \tau^2_{rel})$ from the prior distribution of parameters specifying a Gaussian random field. Typically the prior is specified in the same manner as when calling krige.bayes.

Usage

sample.prior(n, kb.obj=NULL, prior=prior.control())

Arguments

n

number of samples

kb.obj

on object with an output of krige.bayes.

prior

an call to prior.control. Unnecessary if kb.obj is provided.

Value

• A $p+3 \times 4$ data-frame with a sample of the prior distribution of model parameters, where $p$ is the length of the mean parameter $\beta$.

References

Further information on the package geoR can be found at: http://www.est.ufpr.br/geoR.

See Also

krige.bayes and sample.posterior.

Examples

sample.prior(50, prior=prior.control(beta.prior = "normal", beta = .5, beta.var.std=0.1, sigmasq.prior="sc", sigmasq=1.2, df.sigmasq= 2, phi.prior="rec", phi.discrete = seq(0,2, l=21)))