R help - May 2010 - how to get components / factors in factanal / princomp not loadings

Dear all, 

i wonder if there?s a command to obtain the actual values of a principal
component or a factor (not as.factor, but factanal) .


test=princomp(USArrests, cor = TRUE)
summary(test)

just outputs, standard deviation, Prop of Variance and cumulative proportion of
variance.

test$loadings offers yet another proportion of variance scheme. why is that?  

Apart from that:

 Is there a possibily to get the component itself, i mean the linear combination
that represents it vor each observation ? (Loading_i * variable_variable_i)

thx in advance, best regards

matt

test$scores gives you the principal components for princomp

For factanal, you specify eg :
test <- factanal(USArrests,1,scores="regression")
test$scores

see ?factanal

regarding the variance : the "variance" you see is coming from the SS
loadings, which is the Sum of Squared loadings. This divided by the number
of variables (which equals the number of loadings) gives you the proportion
var. this is NOT the proportion of the variance of the original data that
gets explained. In the princomp, it doesn't even make sense as the SS
loadings is bound to be 1, always. It is just displayed as the print
function for the loadings comes from the factor analysis. There the SS
loadings is supposed to be the eigenvalue or latent root associated with
that factor.

Most of this info is available in the help files btw.

Cheers
Joris



On Thu, May 6, 2010 at 5:21 PM, Bunny, lautloscrew.com <
bunny@lautloscrew.com> wrote:
> Dear all,
>
> i wonder if there´s a command to obtain the actual values of a principal
> component or a factor (not as.factor, but factanal) .
>
>
> test=princomp(USArrests, cor = TRUE)
> summary(test)
>
> just outputs, standard deviation, Prop of Variance and cumulative
> proportion of variance.
>
> test$loadings offers yet another proportion of variance scheme. why is
> that?
>
> Apart from that: the "variance" you see is calculated from the SS
loadings,
> being - if I read it right - the sum of squared loadings.
>
>  Is there a possibily to get the component itself, i mean the linear
> combination that represents it vor each observation ? (Loading_i *
> variable_variable_i)
>
> thx in advance, best regards
>
> matt
>
> ______________________________________________
> R-help@r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide
> http://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.
>


-- 
Joris Meys
Statistical Consultant

Ghent University
Faculty of Bioscience Engineering
Department of Applied mathematics, biometrics and process control

Coupure Links 653
B-9000 Gent

tel : +32 9 264 59 87
Joris.Meys@Ugent.be
-------------------------------
Disclaimer : http://helpdesk.ugent.be/e-maildisclaimer.php

	[[alternative HTML version deleted]]