String attributes usually have to get preprocessed before classifiers can work with the data. You can use the StringToWordVector filter for this.
Cheers, Peter
--
Peter Reutemann, Dept. of Computer Science, University of Waikato, NZ
http://www.cms.waikato.ac.nz/~fracpete/ Ph. +64 (7) 858-5174
--
You received this message because you are subscribed to the Google Groups "WekaMOOC-general" group.
To unsubscribe from this group and stop receiving emails from it, send an email to wekamooc-gener...@googlegroups.com.
To post to this group, send email to wekamooc...@googlegroups.com.
Visit this group at http://groups.google.com/group/wekamooc-general.
To view this discussion on the web, visit https://groups.google.com/d/msgid/wekamooc-general/62637a7c-1d52-4856-8af0-cd7b41274a46%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.
Adding to Peter's comment, have you tried this filter with the Filtered Classifier using Bayes, which handles the filtering and classification?
Best wishes, Brooke