R/icproportion.R

In CorReg: Linear Regression Based on Linear Structure Between Variables

icproportion <- function(prop = prop, n = n, alpha = 0.05) {
   inter = alpha / 2 # seuil de confiance
   z = qnorm(1 - inter)
   basic = (2 * n * prop + z^2) / (2 * (n + z^2))
   complement = (z * sqrt(z^2 + 4 * n * prop * (1 - prop))) / (2 * (n + z^2))
   lower = basic - complement
   upper = prop + complement
   return(c(lower, upper))
}