We currently have names of files (e.g. training-data.arff.gz) hard-coded in
Constants for Weka/Meka and Mallet.
As we are adding more ML frameworks this complicated things.
We should introduced some mechanism were proper names are provided by the ML
adapter itself (it should know best).
Original issue reported on code.google.com by torsten....@gmail.com on 9 Sep 2014 at 11:40
Original issue reported on code.google.com by
torsten....@gmail.com
on 9 Sep 2014 at 11:40