BagWhiskerPlot

Implementation for the paper ‘The Bag-and-Whisker Plot: A New Bagplot for Bivariate Data’.

This package exposes a single high-level function:

• bag_whisker(x, y, ...) — compute and (by default) plot a bag-and-whisker plot. x and y are numeric vectors or x may be a two-column matrix/data.frame.
• type1 can be either 'FDR', 'FWER', 'PFER' or 'unadjusted'. q is the control level for the multiple testing procedure.
• When type1 is set to 'unadjusted', the factor lambda is set as the value factor (with default value 3).

Usage example

library(BagWhiskerPlot)

set.seed(1)
n <- 200
clean <- mvrnorm(n, mu = c(0, 0), Sigma = matrix(c(1, 0.6, 0.6, 1), 2))

contam_frac1 <- 0.10
k1 <- floor(n * contam_frac1)
contam1 <- cbind(rnorm(k1, 6, 1), rnorm(k1, -6, 1))
dat_clean <- clean
dat_cont1 <- clean
dat_cont1[1:k1, ] <- contam1

bag_whisker(dat_clean, type1 = 'FDR', q = 0.1, main = "Clean")
bag_whisker(dat_cont1, type1 = 'FDR', q = 0.1, main = paste0("Contaminated ", contam_frac1 * 100, "%"))