Weka

The University of Waikato in New Zealand released Weka 3.0, a collection of machine learning algorithms rewritten entirely in Java.

Why it matters

Classical methods could be applied without programming them, and machine learning reached beyond people who write code.

Weka grew out of a project funded by the New Zealand government from 1993: the first public release (2.1, October 1996) ran on TCL/TK, and in the middle of 1999 Weka 3.0 came out, 100 per cent Java and without a graphical interface; it accompanied the first edition of Witten and Frank's book. The three graphical interfaces appeared between 3.0 and 3.4 (November 2003). By a 2009 paper the package had been downloaded more than 1.4 million times since April 2000, and is accepted in both academia and business. The name is a local flightless bird. The project is also telling in another way: an influential tool came from far outside the large centres.

Event record

Event date

1999 · Approximate date

Timeline date

Event date

Verification

Sources gathered automatically · September 17, 2026

Lines

Tools

ID

evt-0246

The Java rewrite, Weka 3.0, was released in the middle of 1999; the project began in 1993.

Sources

primary Weka 3: Machine Learning Software in Java
University of Waikato
primary The WEKA Data Mining Software: An Update (Hall, Frank, Holmes, Pfahringer, Reutemann, Witten)
ACM SIGKDD Explorations (copy at University of Waikato Research Commons) · Published 2009
secondary Weka (software)
Wikipedia

Related events

Earlier

Related Induction of Decision Trees 1986
Decision trees and other classical methods became available as ready implementations.

Later

Related Torch 2002links to this record
A second attempt to gather machine learning methods into a generally available library.
Related scikit-learn 2010links to this record
The same idea of ready implementations, but as a library for the language scientists write in.