Weka Categorization Commands

logo

it’s known that WEKA crashes when the input dataset is too big. For this reason you have to run the algorithms from your terminal avoiding the GUI. Because the length of the commands is too big, I developed this script which takes as input the installation directory of WEKA, the directory of dataset and the output directory. Then a menu with available algorithms will appeared and you have to choose one. Finally when the algorithm has terminated the results is visible in the terminal.

Built and run

$ git clone https://github.com/sp1thas/WEKACMDs.git && cd WEKACMDs
$ python WekaCommands.py

Prerequisites

  • Python 2.7
  • termcolor

Installation (run as root):

$ pip install -r requirements.txt
  • WEKA link
  • Your dataset

Usage


$ python WekaCommands.py -i <inputfile> -o <outputfile> -w <wekadirectory>

  -i, --ifile

          This is the input dataset

  -o, --ofile
          This is the output file with classification results
          (model is not contained)

  -w, --wekadir
          Direction with WEKA software

  -h,
          Prints these options

Demo

Algorithms Availability

BayesAvailability
BayesNet
NaiveBayes
NaiveBayesMultinomial
NaiveBayesMultinomialText
NaiveBayesUpdateable
FunctionsAvailability
Logistic
MultilayerPerceptron
SimpleLogistic
SMO
LazyAvailability
IBk
KStar
LWL
MetaAvailability
AdaBoostM1
AdditiveRegression
AttributeSelectedClassifier
Bagging
ClassificationViaRegression
CostSensitiveClassifier
CVParameterSelection
FilteredClassifier
IterativeClassifierOptimizer
LogitBoost
MultiClassClassifier
MultiClassClassifierUpdateable
MultiScheme
RandomCommittee
RandomizableFilteredClassifier
RandomSubSpace
RegressionByDiscretization
Stacking
Vote
WeightedistancesHandlerWrapper
MiscAvailability
InputMappedClassifier
SerializedClaassifier
RulesAvailability
DecisionTable
JRip
M5Rules
OneR
PART
ZeroR
TreesAvailability
DecisionTableDecisionStump
HoeffdingTree
J48
LMT
M5P
RandomForest
RandomTree
REPTree
TreesAvailability
RBFNetwork

Authors

  • Simakis PanagiotisInitial work

License

This project is licensed under the GNU General Public License version 3 – see the LICENSE file for details