This is a naive attempt. I am working on a sample-gene microarray dataset with 12 samples (6 normal and corresponding 6 tumor) and 356 EST values. How can I move forward with building an SVM classifier? The data is made available to me as a data frame in R. Also, is it a one-one classification or a one-all classification problem? Should I proceed with building a classifier for each (normal-tumor) pair columns and continue with all 6 such pairs to finally average out a result?