[R] SVM

[Steve Lianoglou]

> Thank you again for your reply! What I would like to do is to class  
> my sample into two group (0,1). I am not sure which method to apply  
> and whether the svm is the correct one! However, when I apply the  
> below R code I get two group the samples in TRUE and FALSE. Can I  
> take this result to put TRUE samples in group A and FALSE samples in  
> group B?

The SVM packages in R implement SVMs that need to be trained in a  
"supervised" setting, which means that you have to have data *with*  
labels. Is this what you have?

The SVM will "look" at the features of each data point and its label,  
and return you a model that can look at *similar* data and predict its  
label (+1 or -1). Since you *already know* the labels of your data  
points, you can give the SVM some of your data that it didn't use for  
training and see how well it can predict their labels.

Lastly, I'm not sure where FALSE and TRUE are coming from, sorry.

> PS: I will not give up to understand statistics!

You shouldn't give up, but unfortunately a mailing list isn't the best  
place to learn statistics (or anything really) from.

If you're interested in really learning more about SVMs and other  
machine learning approaches, you can watch Andrew Ng's intro to  
machine learning classes online. The relevant links are at the top of  
this page:

http://ai.stanford.edu/~ang/courses.html

-steve

--
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