miaSpectra2D {ChemoSpec2D}

Multivariate Image Analysis (Tucker 1) of a Spectra2D Object

Description

Carry out multivariate image analysis of a Spectra2D object (multivariate image analysis is the same as a Tucker1 analysis). Function pcasup1 from package ThreeWay is used.

Usage

miaSpectra2D(spectra)

Arguments

spectra

An object of S3 class Spectra2D.

Value

A list per pcasup1. Of particular interest are the elements C containing the eigenvectors and 1c containing the eigenvalues. We add the class mia to the list for our use later, as well as a method element for annotating plots.

Author(s)

Bryan A. Hanson, DePauw University.

References

A. Smilde, R. Bro and P. Geladi "Multi-way Analysis: Applications in the Chemical Sciences" Wiley (2004). See especially Example 4.5.

P. Geladi and H. Grahn "Multivariate Image Analysis" Wiley (1996). Note that in this text the meanings of scores and loadings are reversed from the usual spectroscopic uses of the terms.

See Also

For other data reduction methods for Spectra2D objects, see pfacSpectra2D and popSpectra2D.

Examples

library("ggplot2")
data(MUD1)
res <- miaSpectra2D(MUD1)

# plotScores & plotScree use ggplot2 graphics

p1 <- plotScores(MUD1, res, tol = 1.0, ellipse = "cls")
p1 <- p1 + ggtitle("MIA Scores")
p1

p2 <- plotScree(res)
p2

# plotLoadings2D uses base graphics
MUD1a <- plotLoadings2D(MUD1, res,
  load_lvls = seq(-90, 0, 10),
  main = "MIA Comp. 1 Loadings"
)

# Selection of loading matrix levels can be aided by the following
# Use MUD1a$names to find the index of the loadings

inspectLvls(MUD1a,
  which = 11, ylim = c(0, 80),
  main = "Histogram of Loadings Matrix"
)