[kepler-dev] Weka actors - experimental release

[Peter Reutemann]

Hi everybody

Since a few of my modifications regarding the Ptolemy II framework got
committed, I spent the last two weeks cleaning up my code and adding
new actors. Currently, most of the Weka functionality (as far as I can
tell) is included:
- associators
- classifiers
- clusterers
- filters
- data generators
- evaluation: random split, cross-validation, testset-based
- loading from/saving to files and databases
- loading/saving of models (classifiers and clusterers at the moment)
- batch and incremental training of classifiers/clusterers
- visualization: scatter plots, attribute histograms, ROC and cost
curves, tree/graph visualizer

You can download the Weka actors from the following URL, section "Kepler":
  http://www.scms.waikato.ac.nz/~fracpete/downloads/

After extracting the archive, just follow the instructions listed in
the file INSTALL. Example datasets and workflows are provided as well.
Even though the archive contains sources (released under Gnu GPL 2) as
well next to the binaries, there's no need to compile the sources
yourself. But there's still a section in the INSTALL file that covers
compilation.

Please keep in mind, that this is still an early development stage.
Feedback welcome! :-)

Cheers, Peter
-- 
Peter Reutemann, Dept. of Computer Science, University of Waikato, NZ
http://www.cs.waikato.ac.nz/~fracpete/ Ph. +64 (7) 858-5174