This will provide 100(1-a)% simultaneous confidence bands for a
sample of size n
. It does this by the "tail-sensitive" approach
of Aldor-Noiman et al (2013), which uses simulated uniform vectors. The
number of simulations is controlled by nsamp
.
ts_bands(n, nsamp = 1000, a = 0.05)
Sample size.
Number of simulation repetitions.
The significance level.
A list of length 3. The $lower
and $upper
confidence
limits at uniform quantiles $q
.
The procedure used is described in Aldor-Noiman et al (2013). But note that they have a mistake in their paper. Step (e) of their algorithm on page 254 should be the CDF of the Beta distribution, not the quantile function.
Aldor-Noiman, S., Brown, L. D., Buja, A., Rolke, W., & Stine, R. A. (2013). The power to see: A new graphical test of normality. The American Statistician, 67(4), 249-260.