PCA - Principal Component Analysis, an alternative?

["Berntsson, Martin" <martin berntsson eon se>]

Dear All,

 

Im working with data science and one of my favorites is PCA - Principal Component Analysis.

Gnumerics is really great in many ways, but the present PCA has some weaknesses.

I found the R library “mixOmics” which has a great PCA (and also PLS).

This supports (among other things) :

·        missing values

·        all matrix forms

·        scaling/centering

·        NIPALS and SVD

 

Suggestion : Can the present PCA be replaced with the one in R “mixOmics” ?

 

If this is of interest/possible, I would gladly help with suggestions around implementation etc.

 

Thanks for a Great software

 

Martin