Weka
Hamilton, New Zealand
What it is
Weka is open-source machine-learning software released under the GNU General Public License. Weka 3.8 is the stable version, while 3.9 is the development version and may contain changes that break compatibility. The stable branch receives bug fixes and feature upgrades intended to preserve compatibility with earlier releases. Weka provides a Java API and can also be used through Jupyter notebooks; the latest official releases require Java 8 or later.
Downloads are available for Windows, Mac OS, and Linux, with Intel and ARM options. Both 3.8 and 3.9 have a package manager for community-added functionality, but downloading and installing packages requires an internet connection. The software includes built-in help and a comprehensive manual, with documentation, mailing list archives, forums, and a mailing list available for questions and bug reports. Weka is free. Distributed derivative works must use the GPL; an appropriate license may be available for commercial projects that need to distribute Weka code in a non-GPL program. Serialized models from version 3.7 do not work with 3.8, and the documented migrator has an exception for RandomForest.
For and against
For
- Free and open-source under the GPL.
- Provides a Java API and Jupyter notebook use.
- Downloads support Windows, Mac OS, and Linux.
- Includes built-in help and a comprehensive manual.
Against
- Installing packages requires an internet connection.
- Latest official releases require Java 8 or later.
- Version 3.7 serialized models are incompatible with 3.8.
The verdict
Weka offers a free machine-learning toolkit across desktop platforms, with stable and development releases. Check version compatibility and GPL requirements before adopting it in a distributed project.
Who it suits
Weka may suit users who want free, open-source machine-learning software with a Java API or notebook workflow. Developers who distribute Weka code in a non-GPL program should review the licensing terms.
Plans and prices
Weka open-source software
Desktop Java workbench · stable and development branches · package manager requires internet
Read 15 Oct 2026 · weka.waikato.ac.nz
Weka 3.8 Stable
Open-source software under the GNU General Public License
Read 3 Oct 2026 · weka.waikato.ac.nz
Weka 3.9 Development
Development version; may include features that break compatibility with earlier releases
Read 3 Oct 2026 · waikato.github.io
On the record
- Association rules
- Weka contains association-rule mining functionality for discovering relationships in transactional data.waikato.ac.nz · read 15 Oct 2026
- Classification
- Weka includes algorithms for assigning records to classes.waikato.ac.nz · read 15 Oct 2026
- Clustering
- The workbench supports unsupervised clustering analysis.waikato.ac.nz · read 15 Oct 2026
- Commercial licensing
- Commercial redistribution that cannot comply with GPL may require a separate licence from relevant copyright holders.waikato.github.io · read 15 Oct 2026
- Community support
- The project directs users to documentation, mailing list archives, community forums, and a mailing list for help and bug reports.waikato.github.io · read 3 Oct 2026
- Compatibility limit
- Serialized Weka models created in version 3.7 are incompatible with version 3.8, and the documented migrator has at least one exception: RandomForest.waikato.github.io · read 3 Oct 2026
- Database integration
- Weka can work with transactional database data through database connectivity and JDBC drivers.waikato.github.io · read 15 Oct 2026
- Data preparation
- The software provides tools for preparing and transforming datasets before modeling.waikato.ac.nz · read 15 Oct 2026
- Desktop platforms
- Downloads are listed for Windows, Mac OS, and Linux, with Intel and ARM options for each.waikato.github.io · read 3 Oct 2026
- Documentation depth
- Documentation includes a built-in manual, free book appendix, Javadocs, package documentation, videos, slides and a long-running mailing-list archive.waikato.github.io · read 15 Oct 2026
- Extensibility
- New classifiers and filters are implemented by deriving classes from Weka classifier and filter base classes, and packages can expose them in the GUI.waikato.github.io · read 15 Oct 2026
- History
- Weka development began in 1993 at the University of Waikato and its first public release was in 1996.waikato.ac.nz · read 15 Oct 2026
- Included documentation
- Weka includes built-in help and a comprehensive manual.waikato.github.io · read 3 Oct 2026
- Java API
- The distribution exposes Java classes that developers can call from their own programs, with Javadoc and API articles available.waikato.github.io · read 15 Oct 2026
- Java requirement
- The latest official Weka releases require Java 8 or later.waikato.github.io · read 3 Oct 2026
- Jupyter integration
- Official documentation describes using Weka through Jupyter notebooks.waikato.github.io · read 15 Oct 2026
- Licence
- Weka is GPL software; Weka 3.6 uses GPLv2 while versions later than 3.7.5 use GPLv3.waikato.github.io · read 15 Oct 2026
- Limit: documentation and integration
- Software Advice reviewers mention limited documentation, a learning curve, an uncommon file format and difficult Python integration.softwareadvice.com · read 15 Oct 2026
