R help - Mar 2012 - SVM works slowly with 1M observations

Dear All,

I m just curious if this is okay that SVM takes several hours to build the model
based on training data set with ~1 million observations and 7 columns (6
variables + 1 class variable). Here is my code:

model<-svm(result ~ ., data=trainset)

where 'trainset' has 997594 obs. and 7 variables.

Is/are there any ways to improve performance?

Appreciate for any advices.


Best,
-Alex

	[[alternative HTML version deleted]]