[R] raw image data in R and ipred

[Mutlu Ozdogan]

Dear R-help list,

I am very new to R software.  I am trying to use the IPRED (bagging with 
regression trees) package to classify multi-layer raw binary images in R.

I first begin with extracting training data from my binary images.  I do this by 
selecting 5000 random pixels from a 25 layer image (each layer is 1200x1200 
pixels and there are 25 layers all short integer binary) and reading out the 
ASCII file.  The training and pruning in IPRED is done with this data set and 
there are no problems here.

My goal is to apply the trained and pruned trees to the full raw binary image.  
This is where my question is:  

Is it possible to bring a 25 layer binary image data into R, keep it in 
binary format, and apply bagging tree rules (generated with training sample) to 
generate a single layer binary image data containing predicted outcomes?

Thank you very much for your help.

Best,

Mutlu Ozdogan



-------------------------------
Mr. Mutlu Ozdogan
Center for Remote Sensing
Boston University
725 Commonwealth Avenue
Boston, MA 02215 USA
phone (617)353-5981
fax (617)353-3200
ozdogan at bu.edu
-------------------------------