asymcnfB obtains asymmetric bayesian confidence bands on a target function

asymcnfB(DF, DFmat, alpha = 0.05, scale = FALSE)

Arguments

DF

the target distribution/quantile function as a vector

DFmat

the matrix of draws of the distribution, rows correspond to elements in DF

alpha

level such that 1-alpha is the desired probability of coverage

scale

logical for scaling using the inter-quartile range

Value

a list containing

  • cmin - constant to subtract from DF for the lower confidence band

  • cmax - constant to add to DF for the upper confidence band

Examples

set.seed(14); m=matrix(rbeta(500,1,4),nrow = 5) + 1:5 DF = apply(m,1,mean); plot(1:5,DF,type="l",ylim = c(min(m),max(m)), xlab = "Index")
asyCB<- asymcnfB(DF,DFmat = m) lines(1:5,DF-asyCB$cmin,lty=2); lines(1:5,DF+asyCB$cmax,lty=2)