PDA

View Full Version : i need help



m_kamal1976
06-29-2008, 12:54 PM
Dear/all
i am making a research in data mining i want to know how to begin with WEKA to compare come classifiers accuracy and processing time over some of datasets .
i want the easiest step to know how to do that.
i want to know the names of these classifiers in WEKA

c 4.5
CBA
bagged and boosted versions of c 4.5
SJEP (strong jumping emerging pattern)
JEP( jumping emerging pattern)
NEP (noise tolerent emerging patterns)
GNEP (generalized noise tolerent emerging patterns)i want to know if all the experiments will be in experimenter or should i do for these data sets before working on it
my test belong to t-test or wilinxton test
:) hope to find support
thx in advance

Mark
06-29-2008, 05:41 PM
Hi,

C4.5 (release 8) is implemented as weka.classifiers.trees.J48.

weka.associations.Apriori has a option to enable class association rule mining via the CBA algorithm

Bagging and boosting (AdaBoost M1, LogitBoost) can be found in the weka.classifiers.meta package.

Weka doesn't have any implementations of the emerging patterns algorithms.

For large-scale classification experiments (and for statistical t-tests) I'd recommend using the Experimenter.

Cheers,
Mark.