MOCCA
====
MOCCA is short for Monte Carlo Clustering Algorithm. It is an evolution of the DOC clustering algorithm. My work for is to incorporate principle component analysis into the algorithm to detect correlated clusters. This implmenetation is built on top of Opensubspace, which in turn is built on top of Weka.

DIRECTORIES
===========
data - data sets, and expected results
mfiles - Matlab/Octave files
misc - files not easily categoried
workspace - eclipse workspace


SNIPPETS
========
Example of making it run on windows:


**** MOCCA (ABSOLUTE PATHS) ****
  java.exe -classpath "C:\Users\ahoffer\Documents\GitHub\sepc\workspace\OpenSubspace;C:\Users\ahoffer\Documents\GitHub\sepc\workspace\OpenSubspace\lib\*" weka.subspaceClusterer.Mocca -m 10000 -w 01 -i 0.3 -s 0.95 -b 0.35 -g 0 -M F1Measure:Accuracy -t breast.arff -c last

--Notice that the directory which contains the jars ("OpenSubspace\lib\*") ends in an asterisk to indicate all the jar files should be included.
--The .class files are in the directory "OpenSubspace".
--Presumes the file breast.arff resides in the curerent directory.

**** MOCCA (RELATIVE PATHS) ****
java -cp ".;.\lib\*" weka.subspaceClusterer.Mocca -m 5000 -w 1 -i 0.3 -b 0.35 -g 0.2 -t breast.arff -c last


**** MOCCA WITH SUBPACE CLUSTER EVALUATION ****
java -cp ".:lib/*" weka.subspaceClusterer.MySubspaceClusterEvaluation -sc Mocca -t "../../data/breast.arff" -c last -g 0 -w 0.5 -i 0.8 -s 0.95 -maxiter 1000 -a 0.1 -e 0.5 -b 0.35 -label 1 -M F1Measure:Accuracy -path "."

**** TESTRUNNER ****
java -cp ".:lib/*" weka.subspaceClusterer.TestRunner

**** COMPILE ON LINUX ****
javac -cp ".:lib/*" weka/subspaceClusterer/*.java