[R] ksvm question -- help! line search failed...

[Steve Lianoglou]

I guess this is going to be hard to debug w/o knowing more about your  
data.

How big is your dataset?
   * How many observations?
   * How many predictors/features per observation?
How many classes are in the data?
Can you make a small reproducible example?
Does your now cleaned data work w/ e1071?
Does ksvm work if you keep prob.model=FALSE?


On Jul 7, 2009, at 10:23 PM, Michael wrote:

> I got the data working, but now I got another problem with KSVM:
>
> line search fails -2.793708 -0.5831701 1.870406e-05 -5.728611e-06
> -5.059796e-08 -3.761822e-08 -7.308871e-13Error in
> prob.model(object)[[p]]$A :
>  $ operator is invalid for atomic vectors
>
>
>
>
> On Tue, Jul 7, 2009 at 6:45 PM, Steve
> Lianoglou<mailinglist.honeypot at gmail.com> wrote:
>> Hi,
>>
>> On Jul 7, 2009, at 6:44 PM, Michael wrote:
>>
>>> What's wrong? Very sad about this...
>>>
>>> model <- ksvm(x=mytraindata[, -1], y=factor(mytraindata[, 1]),
>>> prob.model=T)
>>> Error in .local(x, ...) : x and y don't match.
>>
>> Same problem you're having with the svm in e1071: I bet your data  
>> is wonky.
>>
>> split your data into the x and y vars first and inspect it to make  
>> sure that
>> they're "clean."
>>
>> -steve
>>
>> --
>> Steve Lianoglou
>> Graduate Student: Physiology, Biophysics and Systems Biology
>> Weill Medical College of Cornell University
>>
>> Contact Info: http://cbio.mskcc.org/~lianos
>>
>>
>>
>>

--
Steve Lianoglou
Graduate Student: Physiology, Biophysics and Systems Biology
Weill Medical College of Cornell University

Contact Info: http://cbio.mskcc.org/~lianos