R help - Apr 2008 - Test its performace on new data of a established neural network or SVM model

Hi:

I have already using the MLInterface to train and build a neural network
model and SVM model to clustering a set of data.  Now I have a new set of
data.  I need use these established NN model and SVM to cluster these new
data without retraining it.  I am aware that I can use predict command to do
this in a linear regressing model.  Should I use predict command to do it?
Is there any other way to do this?


Weihsin Wang
Department of bioinformatics
ITRI, Taiwan


	[[alternative HTML version deleted]]