Attribute Clustering, v2.0.20110317 (code by Michael Mampaey)
http://www.adrem.ua.ac.be/implementations



INSTALLATION:

The source code requires mpfr, available at:
    http://www.mpfr.org/
and pugixml, available at:
    http://pugixml.org/

To compile, simply type:
    make



FILE FORMATS:

accepted file formats are binary 'fimi' files, with extension '.dat' or '.fimi', 
and categorical csv files with '.data' extension.



USAGE:

-f      --file                  data filename
-t      --type                  type of score [0=entropy, 1=twopart MDL, 2=semi-refined MDL, 3=refined MDL, 4=BIC, 5=AIC; default=1]
-s      --search                search strategy [1=hierarchical, 2=anneal, 3=beam search; default=1]
-k      --k                     number of clusters to stop at [default=1]
-x      --stop                  stop when in a local minimum
-b      --beamwidth             beam width [default=5]
-i      --iterations            number of steps in simulated annealing [default=1000]
-l      --lambda                simulated annealing scaling factor [default=1]
-c      --cycles                simulated annealing nr of cycles [default=10]
-p      --partition             compute score of partition in file
-o      --output                final partition output filename
-a      --scores                store all intermediate scores in a file
-v      --verbose               verbosity level [default=1]
-q      --quiet                 quiet mode [-v 0]
