plot.msPCA: Plot Method for msPCA Objects

In ssMRCD: Robust Estimators for Multi-Group and Spatial Data

Plot Method for msPCA Objects

Description

Generic plotting interface for objects of class "msPCA". Depending on the type argument, this function visualizes loadings, scores, screeplots, biplots, or score distances.

Usage

## S3 method for class 'msPCA'
plot(x, type = c("loadings"), ...)

Arguments

x	An object of class "msPCA".
type	Type of plot to produce. One of: "loadings" (default), "screeplot", "scores", "biplot", or "score_distances".
...	Additional arguments passed to internal functions.

Details

If type = "loadings" a heatmap of loadings across groups is displayed. Optional input arguments are:

k	Integer. The k-th principal component will be plotted.
text	Boolean, whether the loading values should also be displayed as text. Default FALSE.
size	Numeric. Text size when text is TRUE.
gnames	Character vector. Names for groups (shown in plots).
cnames	Character vector. Names for variables (shown in plots).
textrotate	Rotation angle of text in the heatmap.
tolerance	Numeric, default 1e-04. Specifies the band of white color values around zero.

If type = "screeplot" boxplots of the explained variance per component and cumulative variance per group are plotted. Optional input arguments are:

text	Boolean, whether the loading values should also be displayed as text. Default TRUE.
size	Numeric. Text size when text is TRUE.
gnames	Character vector. Names for groups (shown in plots).
cutoff	Scalar with default value 0.8. The cumulative percentage cutoff value for the overall explained variance.
textrotate	Rotation angle of text in the heatmap.

If type = "biplot" the loadings are visualized in the first and second component over all groups. Optional input arguments are:

color	Character. Either "variable" (default) when the color should be connected to the variables or "groups" if the color should correspond to the groups.
size	Numeric. Text size when text is TRUE.
alpha	Alpha value for the loading points, default is 0.7 .

If type = "scores" a histogram of the k-th scores per group are shown. Optional input arguments are:

k	Integer. The k-th principal component will be plotted. Default value is one.
ssMRCD An object of class ssMRCD including list elements MRCDcov, MRCDmu, mX. Alternatively, X, groups, mu, Sigma can be provided.. X, groups, mu, Sigma The original data matrix, a vector of groups as for ssMRCD, a list of location vectors and a list of covariance matrices.
Alternatively, a ssMRCD object can be given.
alpha	Alpha value for histogramm, default is 0.7 .

If type = "score_distances" a distance-distance plot of score and orthogonal distances is shown for each group. of the k-th scores per group are shown. Optional input arguments are:

k	Integer. Using the first k PCs for the distances. Default is the number of provided PCs. ssMRCD An object of class ssMRCD including list elements MRCDcov, MRCDmu, mX. Alternatively, X, groups, mu, Sigma can be provided.. X, groups, mu, Sigma The original data matrix, a vector of groups as for ssMRCD, a list of location vectors and a list of covariance matrices.
Alternatively, a ssMRCD object can be given.
shape	Point shape.
size	Numeric. Point size.
alpha	Alpha value for the points, default is 0.7 .

Value

A ggplot2 object or list of plots, depending on the type.

See Also

msPCA, align, screeplot.msPCA, , biplot.msPCA, scores,

Examples

# set seed
set.seed(236)

# create data and setup
data = matrix(rnorm(2000), ncol = 4)
groups = sample(1:10, 500, replace = TRUE)
W = time_weights(N = 10, c(3,2,1))

# calculate covariances
covs = ssMRCD(data, groups = groups, weights = W, lambda = 0.3)

# calculate sparse PCA
pca = msPCA(eta = 1.3, gamma = 0.7, COVS = covs$MRCDcov)

# plot screeplot
plot(x = pca, type = "screeplot")

# align and plot loadings
pca$PC = align(PC = pca$PC, type = "mean")
plot(x = pca, type = "loadings", k = 1)
pca$PC = align(PC = pca$PC, type = "maxvar")
plot(x = pca, type = "loadings", k = 1)
pca$PC = align(PC = pca$PC, type = "largest")
plot(x = pca, type = "loadings", k = 1)

# plot different PCA plots
plot(x = pca, type = "score_distances", k = 2,
groups = groups, X = data, mu = covs$MRCDmu, Sigma = covs$MRCDcov)
plot(x = pca, type = "biplot", color = "variable")
plot(x = pca, type = "scores", ssMRCD = covs, k = 1)
plot(x = pca, type = "loadings", k = 1)

ssMRCD documentation built on Nov. 5, 2025, 7:44 p.m.