sprm (version 1.1)

biplot.sprm: Biplot for sprm objects

Description

This biplot for sprm objects visualizes the original variables which contribute to the model and their impact on the latent components as well as the position of the observations in the transformed space. The data is projected onto two of the latent components.

Usage

## S3 method for class 'sprm':
biplot(x, comps = c(1, 2), 
colors = list(scores = "#0000AA", loadings = "red", background = "#BBBBEE"), 
textsize = 6, arrowshapes = c(25, 0.03), ...)

Arguments

x

object of class sprm.

comps

vector with two integers, referring to the components to be plotted.

colors

list of three elements named scores, loadings and background with color codes or names.

textsize

the text size in which to print the scores and loading names.

arrowshapes

vector of length two containing the angle of the arrowheads and their relative length in npc.

...

further arguments. Currently not used.

Details

The sparsity of the biplot is inherited by the sparsity of the model. Only the contributing variables are included in the plot, which leads to better visualization and easier interpretation.

References

Sven Serneels et al. (2014) Sparse partial robust M regression

See Also

plot.sprm, sprm, biplot.prm

Examples

set.seed(5023)
U1 <- c(rep(3,20), rep(4,30))
U2 <- rep(3.5,50)
X1 <- replicate(5, U1+rnorm(50))
X2 <- replicate(20, U2+rnorm(50))
X <- cbind(X1,X2)
beta <- c(rep(1, 5), rep(0,20))
e <- c(rnorm(45,0,1.5),rnorm(5,-20,1))
y <- X%*%beta + e
d <- as.data.frame(X)
d$y <- y
mod <- sprms(y~., data=d, a=2, eta=0.5, fun="Hampel")
biplot(mod, comps = c(1, 2))