Re: libSVM support in Octave

[Alois Schloegl]

John W. Eaton wrote:

> On  8-Oct-2006, Bill Denney wrote:
>
> | doesn't exist.  Does it still fail if you put a file called transpose.m 
> | in the path that just does this:
> | 
> | function x = transpose(x)
> |   x = x';
> | endfunction
> | 
> | If that works, then we probably need to add a function like that into 
> | the octave distribution for compatibility.
>  

Thanks, this is doing the job.

> I think Bill's solution should work, but for transpose, I think you
> want to use the .' operator.  The ' operator is also known as
> ctranspose in Matlab.
>  
In the particular case it does not matter because it is a real matrix; 
but it in general you are right.

> Yes, we should probably make all these operators available with named
> functions as well.
>  
I agree.

> jwe
>  

This allows to use SVMTRAIN. However, at run-time the function 
SVMPREDICT results in this error.

error: octave_base_value::nzmax (): wrong type argument `matrix'
panic: Segmentation fault -- stopping myself...
attempting to save variables to `octave-core'...
save to `octave-core' complete
Segmentation fault

After some debugging, the error massage is caused in svm_model_matlab.c 
(line 6681) by this command
  [num_samples] = mxGetNzmax(rhs[id]);

The segmentation fault occures only a few lines later, perhaps because 
num_samples is -1.

Again, I'd like asking you whether mxGetNzmax is implemented/supported, 
or not.


Best wishes,
    Alois