[RsR] robust PCA in robComposition

[Diman Todorov]

I am not sure PCA is the best approach when your data is not normally
distributed. There are approaches to discrete PCA but ultimately you may
want to consider alternative methods depending on what it is you are
looking for

Diman

On Thu, 7 Sep 2017 at 06:21 Alice Domalik <adomalik using sfu.ca> wrote:

> My data set includes lots of zeros and 100's.
> I thought that might be the problem, but I could find anything in the
> documentation about requirements for the data frame.
>
> Alice
>
>
> From: "Filzmoser Peter" <peter.filzmoser using tuwien.ac.at>
> To: "Alice Domalik" <adomalik using sfu.ca>, r-sig-robust using r-project.org
> Sent: Wednesday, September 6, 2017 8:26:40 PM
> Subject: Re: robust PCA in robComposition
>
> Hard to say without no more details about the data frame. Maybe there are
> zeros included?
> Best regards,
> Peter
>
>
> ------------------------------------------------------------------------------
> Prof. Dr. Peter Filzmoser
> CSTAT - Computational Statistics
> Institute of Statistics & Mathematical Methods in Economics
> Vienna University of Technology
> Wiedner Hauptstr. 8-10
> 1040 Vienna, Austria
> (DVR: 0005886)
> Tel. +43 1 58801 10560 <+43%201%205880110560>
> FAX +43 1 58801 105699 <+43%201%2058801105699>
> e-mail: P.Filzmoser using tuwien.ac.at
> http://www.statistik.tuwien.ac.at/public/filz/
>
> ------------------------------------------------------------------------------
>
> ________________________________________
> From: R-SIG-Robust <r-sig-robust-bounces using r-project.org> on behalf of
> Alice Domalik <adomalik using sfu.ca>
> Sent: Thursday, September 7, 2017 3:13 AM
> To: r-sig-robust using r-project.org
> Subject: [RsR] robust PCA in robComposition
>
> I am trying to perform a robust PCA analysis on compositional data. My
> data set includes 400 diet samples from a species of predatory fish.
> For each sample, I have the percent of the diet that is comprised of 7
> species of prey item. Each row sums to 100.
>
> sandlance rockfish salmon saury anchovy herring squid
> 100.00000 0.00000 0.00000 0.00000 0 0.000000 0.00000
> 100.00000 0.00000 0.00000 0.00000 0 0.000000 0.00000
> 75.00000 0.00000 0.00000 0.00000 0 25.000000 0.00000
>
> I tried "pcaCoDa(diet.df)" (from the package robComposition), b ut I get
> the following error:
>
>
>
> Error in x.ilr[, i] <- sqrt((i)/(i + 1)) * log(((apply(as.matrix(x[,
> 1:i]), : incorrect number of subscripts on matrix
>
>
>
>
> What does this error mean and how can I resolve the problem?
>
> [[alternative HTML version deleted]]
>
> _______________________________________________
> R-SIG-Robust using r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-sig-robust
>
>         [[alternative HTML version deleted]]
>
> _______________________________________________
> R-SIG-Robust using r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-sig-robust
>

	[[alternative HTML version deleted]]