Hi everybody,
I’ve got a data setwith unbalanced class distribution, which I need to resample for making it balanced. Is there in KNIME anything like the weka.filters.supervised.instance.Resample?
Thank you for your help
Cheers
Carmelo
You may have a look at the SMOTE node. Though not a generic oversampling node it allows at least for synthetic minority oversampling.
See also: http://www.cs.cmu.edu/afs/cs/project/jair/pub/volume16/chawla02a.pdf
The one problem with SMOTE is that depending on the dataset it can scale very poorly.