robustpca: Function that uses robust PCA and estimates robust factors...

In RCTS: Clustering Time Series While Resisting Outliers

Function that uses robust PCA and estimates robust factors and loadings.

Description

Contains call to MacroPCA()

Usage

robustpca(object, number_eigenvectors, KMAX = 20, verbose_robustpca = FALSE)

Arguments

object	input
number_eigenvectors	number of eigenvectors to extract
KMAX	The maximal number of principal components to compute. This is a parameter in cellWise::MacroPCA()
verbose_robustpca	when TRUE, it prints messages: used for testing (requires Matrix-package when set to TRUE)

Details

Notes:

Different values for kmax give different factors, but the product lambdafactor stays constant. Note that this number needs to be big enough, otherwise eigen() will be used. Variation in k does give different results for lambdafactor

MacroPCA() crashes with specific values of dim(object). For example when dim(object) = c(193,27). This is solved with evade_crashes_macropca(), for those problematic dimensions that are already encountered during tests.

Value

list with as the first element the robust factors and as the second element the robust factor loadings

RCTS documentation built on May 31, 2023, 9:20 p.m.