Monika Ray wrote:
>
> Hello,
>
> I know that SVM can deal with high dimensional data...but is there any
> literature as to the max limit ? I mean gene microarray data sometimes
> have 22000 features...will svm be able to deal with such high dimensions?
> Also, even if svm can..would this not also depend on the speed of
> the computer in order to analyse so many features in a reasonable time?
>
> Another question is that SVM is well equipped for binary classification.
> However, for multi-class classification..do you just repeat the SVM on
> each new class? for example if one has 4 classes, one first
> differentiates between what is in class 1 and whats not..and then remove
> the data belonging to class 1 and then differentiate between whats in
> class 2 and all others and so on and so forth?
>
> wouldn't think technique also limit in some way the number of dimensions
> the svm can analyse?
>
> Thank You.
>
> Sincerely,
> Monika Ray
>
> ***********************************************************************
> The sweetest songs are those that tell of our saddest thought...
>
> Computer Science, Washington University, St. Louis, MO, USA
> ***********************************************************************
Hi Monika
So far I know SVM can consider 30,000 attributes. I saw a literature
before.
You don't need to split the data class wise. SVM can handle multiclass
data at the same time. Good Luck.
Sincerely yours.
S.Ali
--------------------------------
Shawkat Ali
School of BusSys
Faculty of IT
Monash University
VIC 3800,Australia
--------------------------------------------
Mobile : 0403 613 791 (AU)
61 403 613 791 (OS)
Phone : 61 03 9905 5353 (w)
Fax : 61 03 9905 5159
E-mail : [log in to unmask]
Home page: http://www.bsys.monash.edu.au/people/shawkat/
|