I don’t think it’s possible to monitor progress this way with the current implementation
of BestFirst search. For a faster alternative, you could try
This is a package for WEKA 3.7 that implements two search methods for wrapper-based subset
selection whose time complexity is linear in the number of attributes (ignoring the time
complexity for the base learner: e.g., for a standard decision tree learner, time
complexity would be quadratic in the number of attributes). For details, see the paper.
You could also try WrapperSubsetEval in conjunction with
which has the same time complexity.
On 6/03/2016, at 7:38 PM, Yauhya Al
I am performing attribute selection and i am using the Wrapper with J48. I tested with 23
attributes and it took about 8 hours to finish. Now my data extend to over 120 features,
so imagine the time needed. The worse thing is not the time but actually not seeing any
progress while waiting.
Is there anyway that i can implement this process in my code or use command lines so that
• The number of datasets needs searched
• See every dataset that have been traversed
• See the result of each dataset tested
I know that my only option if the above is not doable, is to perform unsupervised
learning (preprocessing) to reduce the features and then use the wrapper, but i wanted to
see if its possible or not to keep track of the progress in the wrapper.
Wekalist mailing list
Send posts to: Wekalist(a)list.waikato.ac.nz
List info and subscription status: http://list.waikato.ac.nz/mailman/listinfo/wekalist
List etiquette: http://www.cs.waikato.ac.nz/~ml/weka/mailinglist_etiquette.html