[R] Multiple regressions

[Steve Lianoglou]

Hi,

On Mon, Jun 14, 2010 at 4:24 PM, Amy Hessen <amy_4_5_84 at hotmail.com> wrote:
> Hi Steve,
>
> Thank you very much for your reply.
> Could you please tell me how I can know that a given dataset is non linear
> regression?

This is the type of question you should really consult a (local) stat
expert for. I have some ideas, but I'm loathe to offer them because
the *real* stat gurus here will likely shoot them down due to their
lack of rigor, and will most likely just tell you to consult a local
expert.

> What kind of regression SVM (of package e1071) does?
> Linear, multiple, or non linear?

I think you'd benefit on reading up on svm's and regression. Googling
for "support vector regression" will land you on several tutorials.
Support vector regression is also covered in "The Elements of
Statistical Learning," which is available free online.

We already mentioned SVR can handle multiple regression. I'm not sure
how to really answer your "nonlinear" question. Technically it's doing
linear regression, but recall that you can use a kernel in your SVM
that can transform your feature space in interesting (and nonlinear)
ways. The machine will be doing linear regression in the new feature
space, but this can map to some non linear function in the original
input space.

-- 
Steve Lianoglou
Graduate Student: Computational Systems Biology
 | Memorial Sloan-Kettering Cancer Center
 | Weill Medical College of Cornell University
Contact Info: http://cbio.mskcc.org/~lianos/contact