Thanks for reply,

I have one more question, how it calculate accuracy with 10-fold cross 
correlation?

 

On Tuesday, 28 January 2014 13:32:32 UTC+1, Jordi Inglada wrote:
>
> saygin <[email protected]<javascript:>> 
> wrote: 
> > 
> > Hello, 
> > 
> > I am trying to train my .tif image with approximately 1500 pixels. I 
> have 4 classes and one of them has the lowest training with 115 pixels. If 
> I chose training and validation sample ratio 0.5, and  using SVM OpenCV RBF 
> with parameter 
> > optimization I get 232 training and 184 validation at log file. I am not 
> very familiar with OTB and want to ask does not it use approximately half 
> of all training pixels? 
> > 
>
> No. It will set the number of samples per class to the number of samples 
> of the class which has the fewer samples. In your case 115. If you have 4 
> classes, you get: 
>
> 115 * 0.5 * 4 = 230 
>
> Then, why the number of training samples is not equal to the number of 
> validation? I think it is because first the samples are selected, and then 
> they are randomly split between training and validation, and this may 
> introduce some variation accros classes. 
>
> The current choices for the sampling procedure need improvement and some 
> thinking about this is ongoing (
> http://wiki.orfeo-toolbox.org/index.php/Refactoring_of_the_classification_chain).
>  
> Do not hesitate to give you 2 cents about it. 
>
> Jordi 
>
>

-- 
-- 
Check the OTB FAQ at
http://www.orfeo-toolbox.org/FAQ.html

You received this message because you are subscribed to the Google
Groups "otb-users" group.
To post to this group, send email to [email protected]
To unsubscribe from this group, send email to
[email protected]
For more options, visit this group at
http://groups.google.com/group/otb-users?hl=en
--- 
You received this message because you are subscribed to the Google Groups 
"otb-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
For more options, visit https://groups.google.com/groups/opt_out.

Reply via email to