Re: Convertion from nominal to numeric

Peter Reutemann <[email protected]> Mon, 9 Jan 2023 11:44:09 +1300
Newsgroups gmane.comp.ai.weka
Message-ID <CAHoQ12KZMB9A=tZTzgxy1nzv8EqyowrOzjOLUPEx9LmE9DY26A@mail.gmail.com>
> Yes! I just noticed that. There is a space before each of the attribute values.
> How can I handle that?
> Should there be quotes around the entire expression? If you had that for GB and GK below?
> weka.filters.unsupervised.attribute.RenameNominalValues -R 1 -N " GB:11, GK:12"
> Or quotes around each value?
> weka.filters.unsupervised.attribute.RenameNominalValues -R 1 -N " GB":11," GK":12
> Should they be single quotes or double quotes?

I just had a quick look at the code of the filter and it performs a
"trim" (ie removes leading/trailing white spaces like spaces), so
including spaces in the labels won't work unfortunately.

The simplest approach, if you want to stick with Weka, is to just open
the file in a text editor and replace the leading spaces. If the data
is stored in an ARFF file, then values with spaces should be enclosed
in single quotes, which will make it easy to replace.

If you want to automate these manual preprocessing steps, you might
consider switching to ADAMS (https://adams.cms.waikato.ac.nz/) and its
workflow system. It includes Weka (when downloading an "ml-app"
snapshot or release) and a lot more operators for dealing with
spreadsheets. We had to deal with a lot of data in spreadsheet format,
which resulted in writing a lot of components/operators for dealing
with these things. On top of that, the Weka Investigator that comes
with ADAMS is a more flexible/powerful interface compared to the Weka
Explorer. :-)
(Disclaimer: I'm the lead developer of ADAMS)

Cheers, Peter
-- 
Peter Reutemann
Dept. of Computer Science
University of Waikato, Hamilton, NZ
Mobile +64 22 190 2375
https://www.cs.waikato.ac.nz/~fracpete/
http://www.data-mining.co.nz/
_______________________________________________
Wekalist mailing list -- [email protected]
Send posts to [email protected]
To unsubscribe send an email to [email protected]
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