archetypes: Perform archetypal analysis on a data matrix.

Description

Perform archetypal analysis on a data matrix.

Usage

archetypes(data, k, weights, maxIterations=100,
    minImprovement=sqrt(.Machine$double.eps), maxKappa=1000,
    verbose=FALSE, saveHistory=TRUE,
    family=archetypesFamily("original"), ...)
robustArchetypes(data, k, familyBlocks=list(), ...)
weightedArchetypes(data, k, weights, familyBlocks=list(), ...)

Arguments

data

A numeric $n \times m$ data matrix.

The number of archetypes.

weights

Data weights matrix or vector (used as elements of the diagonal weights matrix).

maxIterations

The maximum number of iterations.

minImprovement

The minimal value of improvement between two iterations.

maxKappa

The limit of kappa to report an ill-ness warning.

verbose

Print some details during execution.

saveHistory

Save each execution step in an environment for further analyses.

family

Blocks defining the underlying problem solving mechanisms; see archetypesFamily.

...

Additional arguments for family blocks.

familyBlocks

Exchange predefined family blocks.

Value

archetypes: An object of class archetypes, see archetypes-class. robustArchetypes: An object of class robustArchetypes and archetypes-class. weightedArchetypes: An object of class weightedArchetypes and archetypes-class.

References

Cutler and Breiman. Archetypal Analysis. Technometrics, 36(4), 1994. 338-348.

Examples

Run this code

data(toy)
a <- archetypes(toy, 3)