Wrong axes limits in PCA distance plot when data contains excluded outliers

[svkucheryavski]

When dataset contains hidden (excluded) outliers, plotResiduals() does not take this into account and computes much larger limits. This code reproduces the problem graphically (see the second plot):

library(mdatools)
set.seed(42)

# generate data and add an outlier
X <- matrix(rnorm(100 * 10), 100, 10)
X[20, ] <- X[20, ] * 10

# make two models, one with outlier and one where outlier is excluded
m1 <- pca(X, 5)
m2 <- pca(X, 5, exclrows = 20)

# show the plots
par(mfrow = c(1, 3))
plotResiduals(m1)
plotResiduals(m2)
plotResiduals(m2, show.excluded = TRUE)

The problem is in ldecomp.plotResiduals() function, which code has to be refactored.