missMDA (version 1.14)

plot.MIPCA: Plot the graphs for the Multiple Imputation in PCA

Description

From the multiple imputed datasets, the function plots graphs for the individuals, variables and dimensions for the Principal Component Analysis (PCA)

Usage

# S3 method for MIPCA
plot(x, choice = "all", axes = c(1, 2), new.plot = TRUE, 
   main = NULL, level.conf = 0.95, …)

Arguments

x

an object of class MIPCA

choice

the graph(s) to plot. By default "all" the graphs are plotted. "ind.proc" the procrustean representation of the individuals, "dim" the representation of the dimensions of the PCA, "ind.supp" the projection of the individuals as supplementary individuals, "var" the projection of the variables as supplementary variables

axes

a length 2 vector specifying the components to plot

new.plot

boolean, if TRUE, a new graphical device is created

main

string corresponding to the title of the graph you draw (by default NULL and a title is chosen)

level.conf

confidence level used to construct the ellipses. By default, 0.95

…

further arguments passed to or from other methods

Value

Four graphs can be drawn:

ind.supp

The individuals of the imputed datasets are projected as supplementary individuals onto the reference PCA map; then confidence ellipses are drawn

var

The variables of the imputed datasets are projected as supplementary variables onto the reference PCA map

ind.proc

A PCA is performed on each imputed dataset and each configuration of scores is rotated onto the reference PCA map with procrustes rotation; then confidence ellipses are drawn

dim

The dimensions of each imputed dataset are projected as supplementary variables onto the dimensions of the reference PCA dimensions

Details

Plots the multiple imputed datasets obtained by the function MIPCA. The idea is to represent the multiple imputed dataset on a reference configuration (the map obtained from the PCA on the incomplete dataset). Different ways are available to take into account and visualize the supplement variability due to missing values.

References

Josse, J., Husson, F. (2010). Multiple Imputation in PCA

See Also

MIPCA,imputePCA

Examples

# NOT RUN {
data(orange)
## nb <- estim_ncpPCA(orange,ncp.max=5) ## Time consuming, nb = 2
resMI <- MIPCA(orange,ncp=2)
plot(resMI)
# }