| credible.region {bayesSurv} | R Documentation |
Compute a simultaneous credible region (rectangle) from a sample for a vector valued parameter.
Description
See references below for more details.
Usage
credible.region(sample, probs=c(0.90, 0.975))
Arguments
sample |
a data frame or matrix with sampled values (one column = one parameter) |
probs |
probabilities for which the credible regions are to be computed |
Value
A list (one component for each confidence region) of length equal to
length(probs). Each component of the list is a matrix with two
rows (lower and upper limit) and as many columns as the number of
parameters giving the confidence region.
Author(s)
Arnošt Komárek arnost.komarek@mff.cuni.cz
References
Besag, J., Green, P., Higdon, D. and Mengersen, K. (1995). Bayesian computation and stochastic systems (with Discussion). Statistical Science, 10, 3 - 66, page 30
Held, L. (2004). Simultaneous inference in risk assessment; a Bayesian perspective In: COMPSTAT 2004, Proceedings in Computational Statistics (J. Antoch, Ed.), 213 - 222, page 214
Held, L. (2004b). Simultaneous posterior probability statements from Monte Carlo output. Journal of Computational and Graphical Statistics, 13, 20 - 35.
Examples
m <- 10000
sample <- data.frame(x1=rnorm(m), x2=rnorm(m), x3=rnorm(m))
probs <- c(0.70, 0.90, 0.95)
CR <- credible.region(sample, probs=probs)
for (kk in 1:length(CR)){
suma <- sum(sample$x1 >= CR[[kk]]["Lower", "x1"] & sample$x1 <= CR[[kk]]["Upper", "x1"] &
sample$x2 >= CR[[kk]]["Lower", "x2"] & sample$x2 <= CR[[kk]]["Upper", "x2"] &
sample$x3 >= CR[[kk]]["Lower", "x3"] & sample$x3 <= CR[[kk]]["Upper", "x3"])
show <- c(suma/m, probs[kk])
names(show) <- c("Empirical", "Desired")
print(show)
}