R/plot_ci.R

In dco61/isteahMAT104: MAT104 Probabilité et Statistiques

#'
#' Fonction plot_ci: tirée de la librairie Openintro
#' Plot confidence intervals
#'
#' @description This function takes vectors of upper and lower confidence
#' interval bounds, as well as population mean, and plots which confidence
#' intervals intersect the mean
#' @param lo a vector of lower bounds for confidence intervals
#' @param hi a vector of upper bounds for confidence intervals
#' @param m the population mean
#' @export
#'
plot_ci <- function(lo, hi, m) {
    nhit <- sum((m >= lo) & (m <= hi))/length(lo)
    par(mar = c(5, 4, 4, 2), mgp = c(2.7, 0.7, 0))
    k <- length(lo)
    ci.max <- max(rowSums(matrix(c(-1 * lo, hi), ncol = 2)))

    xR <- m + ci.max * c(-1, 1)
    yR <- c(0, 41 * k/40)

    plot(xR, yR, type = "n", xlab = "Variable X", ylab = "Échantillon", axes = TRUE)
    abline(v = m, lty = 2, lwd = 2, col = "darkgreen")
    axis(3, at = m, paste("mu = ", round(m, 4)), cex.axis = 1.15)

    # axis(2)
    for (i in 1:k) {
        x <- mean(c(hi[i], lo[i]))
        ci <- c(lo[i], hi[i])
        if ((m < hi[i] & m > lo[i]) == FALSE) {
            col <- "#F05133"
            points(x, i, cex = 1.4, col = col)
            # points(x, i, pch=20, cex=1.2, col=col)
            lines(ci, rep(i, 2), col = col, lwd = 3)
        } else {
            col <- 1
            points(x, i, pch = 20, cex = 1.2, col = col)
            lines(ci, rep(i, 2), col = col)
        }
    }
    mtext(text = paste("Taux de capture = ", nhit * 100, "%"), side = 4, line = 0.5, outer = FALSE,
        cex = 1)
}