R/quantilesLogConDens.r

In logcondens: Estimate a Log-Concave Probability Density from Iid Observations

quantilesLogConDens <- function(ps, res) {
  if (any(ps < 0 | ps > 1)) {
    stop("All entries of the argument ps given to quantilesLogConDens() must be in [0, 1]!\n")
  }

  x <- res$x
  phi <- res$phi
  Fhat <- res$Fhat
  n <- length(x)
  res <- matrix(NA, ncol = 2, nrow = length(ps))

  for (i in 1:length(ps)) {
    p0 <- ps[i]

    if (p0 == 0) {
      q <- -Inf
    }
    if (p0 == 1) {
      q <- x[n]
    }

    if ((p0 > 0) && (p0 < 1)) {
      n <- length(x)
      xj <- max(x[Fhat <= p0])
      j <- length(x[x <= xj])
      q <- xj + (x[j + 1] - x[j]) * qloglin((p0 - Fhat[j]) / (Fhat[j + 1] - Fhat[j]), (x[j + 1] - x[j]) * (phi[j + 1] - phi[j]))
    }

    res[i, ] <- c(p0, as.numeric(q))
  }

  colnames(res) <- c("ps", "quantile")
  return(res)
}