Does Amount of Test Classes matter for Binary Random Forest Classifier

classic Classic list List threaded Threaded
2 messages Options
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Does Amount of Test Classes matter for Binary Random Forest Classifier

micCalve
Hello,

I am curious if there is a cap for the number of classes can be classified/predicted simulatenously using a Binary Random Forest?

I've currently classifying 22 classes at a time, with 250 instances per class and 3 attributes per instance.  I then use a threshold to decide, based upon the number of instances out of 250, that class has been predicted as.  I use this Binary Random Forest as a filter and then whichever ones pass the threshold, get placed into a testing file for a 6 class Multi Random Forest Classification.

Basically, does the number of classes in the testing set matter?  For both a Binary Random Forest, and a MultiClass Random Forest?

Thanks!
Michael

_______________________________________________
Wekalist mailing list
Send posts to: [hidden email]
List info and subscription status: https://list.waikato.ac.nz/mailman/listinfo/wekalist
List etiquette: http://www.cs.waikato.ac.nz/~ml/weka/mailinglist_etiquette.html
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: Does Amount of Test Classes matter for Binary Random Forest Classifier

Eibe Frank-2
Administrator
I think you’ll have to explain more precisely what you are doing. Can you give a step by step description?

Cheers,
Eibe

> On 13/06/2017, at 8:25 AM, Michael Calve <[hidden email]> wrote:
>
> Hello,
>
> I am curious if there is a cap for the number of classes can be classified/predicted simulatenously using a Binary Random Forest?
>
> I've currently classifying 22 classes at a time, with 250 instances per class and 3 attributes per instance.  I then use a threshold to decide, based upon the number of instances out of 250, that class has been predicted as.  I use this Binary Random Forest as a filter and then whichever ones pass the threshold, get placed into a testing file for a 6 class Multi Random Forest Classification.
>
> Basically, does the number of classes in the testing set matter?  For both a Binary Random Forest, and a MultiClass Random Forest?
>
> Thanks!
> Michael
> _______________________________________________
> Wekalist mailing list
> Send posts to: [hidden email]
> List info and subscription status: https://list.waikato.ac.nz/mailman/listinfo/wekalist
> List etiquette: http://www.cs.waikato.ac.nz/~ml/weka/mailinglist_etiquette.html

_______________________________________________
Wekalist mailing list
Send posts to: [hidden email]
List info and subscription status: https://list.waikato.ac.nz/mailman/listinfo/wekalist
List etiquette: http://www.cs.waikato.ac.nz/~ml/weka/mailinglist_etiquette.html
Loading...