[R] Limited number of principal components in PCA
William Armstrong
William.Armstrong at noaa.gov
Fri Jul 29 20:33:07 CEST 2011
Hi all,
I am attempting to run PCA on a matrix (nrow=66, ncol=84) using 'prcomp'
(stats package). My data (referred to as 'Q' in the code below) are
separate river streamflow gaging stations (columns) and peak instantaneous
discharge (rows). I am attempting to use PCA to identify regions of that
vary together.
I am entering the following command:
test_pca_Q<-prcomp(~.,data=Q,scale.=TRUE,retx=FALSE,na.action=na.omit)
It is outputting 54 'standard deviation' numbers (which are the
sqrt(eigenvalues) in respect to a certain PC, am I correct?), and 54
'rotation' numbers, which are the variable loadings with respect to a given
PC.
I have two questions:
1.) Why is it only outputting 54 PCs and standard deviations? If I have 84
variables isn't the maximum number of PCs I can create 84 as well?
2.) Can I now use the 'rotation' values to find clusters of gages that I
acting together, or is there another step I must take?
Thank you very much for your insight.
Billy
--
View this message in context: http://r.789695.n4.nabble.com/Limited-number-of-principal-components-in-PCA-tp3704956p3704956.html
Sent from the R help mailing list archive at Nabble.com.
More information about the R-help
mailing list