2012/1/9 Mathias Verbeke :
> Dear all,
>
> In the documentation of the SVM module, I saw that it was possible to pass
> your own Gram matrix to the kernel. I was wondering if it was also possible
> to do the reverse, i.e. to export the calculated Gram matrix (that gives the
> similarity between the
Dear all,
In the documentation of the SVM module, I saw that it was possible to pass
your own Gram matrix to the kernel. I was wondering if it was also possible
to do the reverse, i.e. to export the calculated Gram matrix (that gives
the similarity between the train and test instances)?
Best and
