WekaRandomForestHelper

WekaRandomForestHelper is a java application that help you to build random forest models and classify some data.

Dependencies

WekaRandomForestHelper uses the jar of Weka which can be found here (The zip archive in the « other platform » section).

It also uses the jar of Jcommander which can be build from the sources found here

Usage

Compile it

cd src
javac *.java -d ../bin -cp ".:path/to/jcommander.jar:path/to/weka.jar"

Run it

java -cp ".:path/to/jcommander.jar:path/to/weka.jar:bin" WekaRandomForestHelper [arguments]

Arguments

There is two command that can be used with this application :

• build
• classify

The build command takes an extra argument : the arff file used to train the model (--data or -d flag)

The classify command takes at least two extra argument :

• the classifier model (--classifier or -c flag)
• at least one arff file of which you want to classify the data

Exemple

(assuming all the required jars are in the classpath)

java WekaRandomForestHelper build --data=resources/weather.nominal.arff

java WekaRandomForestHelper classify --classifier=resources/classifier.model resources/testweather.nominal.arff