"SVM N-grams"

Legacy User

Hi,

I'm a graduate student, and I'm using RapidMiner for some experiments. I'm using several types of SVM, and was asked by my professor to use n-grams SVM.

I tried to find out how to implement this in RapidMiner, but have not been successful. If there is a way to do it, please show me how.

Thanks in advance,

Gil

IngoRM

Hi,

I am not completely sure. If you are in text mining probably calculating n-grams with the text plugin together with a linear kernel should do the trick... however, it could also be structural learning and then: this is not supported yet.

Cheers,
Ingo