Dear Marcel, I'm sorry for the misunderstanding. I think what you really need are the scores values provided by PCA. See ?princomp for details. Here's an example:
# PCA prin.comp=princomp(USArrests, cor = TRUE) summary(prin.comp) # What do you have in prin.comp? names(prin.comp) # Extracting the scores res=prin.comp$scores res # If you want the scores associated to the first component, just type this: res[,1] HTH, Jorge On Thu, Jul 3, 2008 at 8:43 AM, Bio7 <[EMAIL PROTECTED]> wrote: > > Dear Jorge, > > As i've already written the princomp method works for me. But i'm > interested > to produce from the results a matrix which i can visualize (e.g the first > pc) in an image application and which is then the source for a clustering > algorithm. > The background is that i've an application which can transfer images to R > and create images from R very fast (using ImageJ and RServe). I've already > clustered successfully image data with the cluster package. > If i now have more than three channels (or bands - > R,G,B) i want to > reduce > the dataset with the PCA. > So which data (as a matrix) represents the first pc or how can i calculate > from the results > this matrix? Since i've not worked with the PCA method before any > explanation beyond the R help > would really help me. > > > -- > View this message in context: > http://www.nabble.com/PCA-on-image-data-tp18255217p18258404.html > Sent from the R help mailing list archive at Nabble.com. > > ______________________________________________ > 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. > [[alternative HTML version deleted]] ______________________________________________ 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.
