LibD3C: Ensemble classifiers with a clustering and dynamic selection strategy

---

Home

Weka

Data

Contact and Citation

          Install libD3C into weka (Linux environment)

         make sure you have installed libSVM package in your weka!!!

          Environmental requirements:     JDK1.6 or upwards, Weka3.7 or upwards

    (1)Download the latest version of the weka software, and extract file:

(2) Configuration environment (Under the /etc/profile file):

WEKAHOME is the installation path of weka

(4)Install LibD3C package and download classifiers.xml with command:

java weka.core.WekaPackageManager -install-package http://lab.malab.cn/soft/LibD3C/LibD3C.zip

wget http://lab.malab.cn/soft/LibD3C/classifiers.xml

(5)Run LibD3C package with command:

java weka.Run LibD3C -t data/iris.arff -target-correct-rate 1.0 -I 0.05 -validation-ratio 0.2 -ensemble-vote-rule 1 -K 10 -circle-combination-algorithm 3 -selective-algorithm 1 -num-slots 1 -classifiers-xml filePath of your classifiers.xml -time-out 20 -S 1

Tips:

       Currently it is not supported to change only part of the parameters.

       If you want to modify the parameters, you need to pass all the parameters to the classifier.

       We will update it as soon as possible!

(6)Result