Users installing Weka on Linux must have always felt left out a bit, with no installer available, instead having to deal with just a ZIP file. For power users, that would not have mattered, but users new to Linux may have found that a bit more challenging.

Well, things have changed - at least for users of Debian or one of its many derivatives like Ubuntu - with the advent of snapshots being available for download as Debian packages (stable 3.8 and developer version).

Downloading and installing

All snapshots, ZIP files and Debian packages, get generated by our build server and are available from the following URL:

cs.waikato.ac.nz/~ml/weka/snapshots/weka_snapshots.html

Simply download the version that you would like to install on your system, stable or developer.

Once downloaded, you can either install the package from the command-line using sudo dpkg -i filename.deb or double-click it to use a graphical installer like GDebi or Ubuntu's Software Center (for some reason the Software Center thinks that Weka is proprietary).

Starting Weka

Depending on what version you installed, you will have two scripts in your /usr/bin directory:

weka-stable-gui
weka-stable-run

Or:

weka-dev-gui
weka-dev-run

GUI

The -gui script will start Weka's GUIChooser with 512MB of heap space by default. You can increase the heap with the -memory option, e.g., -memory 2g to use 2GB. If you want to execute another GUI class, e.g., the Explorer, then you can supply this with the -main option, e.g., -main weka.gui.explorer.Explorer.

Starting the Explorer of the developer version with 2GB of heap from the command-line would look like this:

weka-dev-gui -main weka.gui.explorer.Explorer -memory 2g

The Debian package also installs a .desktop file in /usr/share/applications, which simply calls the weka-{dev|stable}-gui script and should show up in your applications menu automatically.

Run

The -run script executes Weka's weka.Run class, for executing classes, like classifiers, clusterers, filters, etc. from the command-line. The weka.Run tool allows you to specify shortened classnames (as long as they are unique), e.g., .J48 instead weka.classifiers.trees.J48. This script also offers the -memory option to chage the heap size from its default 512MB.

The following command-line cross-validates J48 (stable version) on the labor UCI dataset:

weka-stable-run .J48 -t /some/where/labor.arff

Building packages

This section is only relevant to people wanting to build their own Debian packages, e.g., when including their own classes in a custom build, or people that want to simply know how it is done.

Building of Debian packages is only possible on Linux systems that have the Debian package tools installed (fakeroot, dpkg-deb), as the build process uses the Debian Maven Plugin to generate them, which in turn relies on these system tools.

There are two different build profiles available, one that uses the Maven build framework to collect all the required artifacts (mvn-deb-pkg) and one that reuses jar files built previously by ant (ant-deb-pkg). Our build server uses the latter one.

Using Maven

You can use the following command-line for building a Debian package (located in your target directory):

mvn clean install -DskipTests=true deb:package -P mvn-deb-pkg

Using Ant

Since the Ant profile relies on previously generated jar files, you need to execute two commands to successfully build a Debian package in your target directory:

ant clean exejar srcjar
mvn deb:package -P ant-deb-pkg