postExp: Exponentiate posterior quantiles

In geostatsp: Geostatistical Modelling with Likelihood and Bayes

Exponentiate posterior quantiles

Description

Converts a summary table for model parameters on the log scale to the natural or exponentiated scale.

Usage

	postExp(x, 
		exclude = grep('^(range|aniso|shape|boxcox)', rownames(x)),
		invLogit=FALSE)

Arguments

x	a matrix or data frame as returned by glgm
exclude	vector of parameters not transformed, defaults to the range parameter
invLogit	Converts intercept parameter to inverse-logit scale when TRUE. Can also be a vector of parameters to inverse-logit transform.

Value

a summary table for log or exponentially transformed model parameters

Examples

require("geostatsp")
data("swissRain")
swissRain = unwrap(swissRain)
swissAltitude = unwrap(swissAltitude)

swissRain$lograin = log(swissRain$rain)

if(requireNamespace('INLA', quietly=TRUE)) {
  INLA::inla.setOption(num.threads=2)
  # not all versions of INLA support blas.num.threads
  try(INLA::inla.setOption(blas.num.threads=2), silent=TRUE)
swissFit =  glgm(formula="lograin", data=swissRain, 
	grid=20, 
	covariates=swissAltitude/1000, family="gaussian", 
	prior = list(sd=1, range=100*1000, sdObs = 2),
	control.inla = list(strategy='gaussian', int.strategy='eb'),
	control.mode = list(theta=c(1.6542995, 0.7137123,2.2404179))
	)
if(length(swissFit$parameters)) {
	postExp(swissFit$parameters$summary)
}

}

geostatsp documentation built on May 29, 2024, 5:53 a.m.