Multiple iterations sampling via GUI

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

Multiple iterations sampling via GUI

David
Hi there,
  I am balancing the training set by for instance using SMOTE, i.e.,
weka.classifiers.meta.FilteredClassifier -F
"weka.filters.supervised.instance.SMOTE -C 0 -K 5 -P 100.0 -S 1" -S 1 -W
weka.classifiers.trees.J48 -- -C 0.25 -M 2

Since sampling requires a random selection of data in the training set, then
it is a common procedure to repeat the (balanced) classification multiple
times (e.g., 100 times). Is there a way to do it in the GUI?




--
Sent from: https://weka.8497.n7.nabble.com/
_______________________________________________
Wekalist mailing list -- [hidden email]
Send posts to: To unsubscribe send an email to [hidden email]
To subscribe, unsubscribe, etc., visit
https://list.waikato.ac.nz/postorius/lists/wekalist.list.waikato.ac.nz
List etiquette: http://www.cs.waikato.ac.nz/~ml/weka/mailinglist_etiquette.html
Reply | Threaded
Open this post in threaded view
|

Re: Multiple iterations sampling via GUI

Eibe Frank-2
Administrator
Unfortunately, SMOTE does not implement the Randomizable interface, otherwise you could just use RandomCommittee with the FilteredClassifier as the base learner. If a filter (or the base classifier) is Randomizable, FilteredClassifier will automatically modify the seed of the random number generator in the base scheme.

The alternative is to shuffle the data before it is passed to SMOTE. One possibility is something like this:

java weka.Run .RandomCommittee -I 100 -W .FilteredClassifier -t ~/datasets/UCI/diabetes.arff -- -F .Randomize -W .FilteredClassifier -- -F .SMOTE -W .J48

Cheers,
Eibe

> On 24/09/2019, at 12:11 AM, David <[hidden email]> wrote:
>
> Hi there,
>  I am balancing the training set by for instance using SMOTE, i.e.,
> weka.classifiers.meta.FilteredClassifier -F
> "weka.filters.supervised.instance.SMOTE -C 0 -K 5 -P 100.0 -S 1" -S 1 -W
> weka.classifiers.trees.J48 -- -C 0.25 -M 2
>
> Since sampling requires a random selection of data in the training set, then
> it is a common procedure to repeat the (balanced) classification multiple
> times (e.g., 100 times). Is there a way to do it in the GUI?
>
>
>
>
> --
> Sent from: https://weka.8497.n7.nabble.com/
> _______________________________________________
> Wekalist mailing list -- [hidden email]
> Send posts to: To unsubscribe send an email to [hidden email]
> To subscribe, unsubscribe, etc., visit
> https://list.waikato.ac.nz/postorius/lists/wekalist.list.waikato.ac.nz
> List etiquette: http://www.cs.waikato.ac.nz/~ml/weka/mailinglist_etiquette.html
_______________________________________________
Wekalist mailing list -- [hidden email]
Send posts to: To unsubscribe send an email to [hidden email]
To subscribe, unsubscribe, etc., visit
https://list.waikato.ac.nz/postorius/lists/wekalist.list.waikato.ac.nz
List etiquette: http://www.cs.waikato.ac.nz/~ml/weka/mailinglist_etiquette.html