Performance of Weka in feature selections?

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

Performance of Weka in feature selections?

asadbtk
In one of the Weka's tutorial, Ian mentioned that Weka can support a maximum of 25 features in a dataset while performing the feature selection algorithms? Is it the case or I have misunderstood it completely? Do a large number of features in a dataset degrade the performance of Weka? In some of my experiments in the recent past, a dataset have had more than 100 features and the Weka's feature selection algorithms (after executing) only selected as few as 5 features. So is it an abnormal case?

Mail priva di virus. www.avast.com

_______________________________________________
Wekalist mailing list
Send posts to: [hidden email]
To subscribe, unsubscribe, etc., visit 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
|

Re: Performance of Weka in feature selections?

Eibe Frank-3
There must have been a misunderstanding. It is certainly true though that feature selection gets harder as the number of features increases.

Cheers,
Eibe

On Fri, Jul 12, 2019 at 8:48 PM javed khan <[hidden email]> wrote:
In one of the Weka's tutorial, Ian mentioned that Weka can support a maximum of 25 features in a dataset while performing the feature selection algorithms? Is it the case or I have misunderstood it completely? Do a large number of features in a dataset degrade the performance of Weka? In some of my experiments in the recent past, a dataset have had more than 100 features and the Weka's feature selection algorithms (after executing) only selected as few as 5 features. So is it an abnormal case?

Mail priva di virus. www.avast.com
_______________________________________________
Wekalist mailing list
Send posts to: [hidden email]
To subscribe, unsubscribe, etc., visit 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]
To subscribe, unsubscribe, etc., visit https://list.waikato.ac.nz/mailman/listinfo/wekalist
List etiquette: http://www.cs.waikato.ac.nz/~ml/weka/mailinglist_etiquette.html