Quick Start and Overview

Pentaho Data Mining, based on Weka project, is a comprehensive set of tools for machine learning and data mining. Its broad suite of classification, regression, association rules, and clustering algorithms can be used to help you understand the business better and also be exploited to improve future performance through predictive analytics.

There are two versions of Weka:

Weka 3.8 - current stable version. This branch receives bug fixes to core Weka; new features are released through packages that can be installed via the built-in package manager.
Weka 3.9 - development branch. This is a continuation of the 3.8 code line that receives both bug fixes and new features/improvements to core Weka. It also takes advantage of new features released in packages.

Documentation

Pentaho Data Mining (Weka)

There is a book that has been written to accompany Weka - Data Mining: Practical Machine Learning Tools and Techniques (Fourth Edition).

Plugins for Pentaho Data Integration (Kettle)

Using the Weka Scoring Plugin (download)
Using the Reservoir Sampling Plugin (included as a first class step in recent Kettle distributions)
Using the ARFF Output Plugin (download)
Using the Univariate Statistics Plugin (included as a first class step in recent Kettle distributions)
Using the Knowledge Flow Plugin (enterprise edition)
Time Series Analysis and Forecasting with Weka (available as a PDI Spoon perspective as well as a Weka plugin)
Weka time series forecasting plugin for PDI 4 (enterprise edition)
3D Visualization Perspective for PDI 4 (download)

Developing with Weka

Awards and Publications

Under Development/Roadmap

Complete Knowledge Flow rewrite - new engine, refactored UI etc.
PMML Support in Weka
Distributed Weka for Hadoop and for Spark
Incremental dictionary creation and vectorisation (StringToWordVector filter) for text documents

Archived

Further Links and Information

Pentaho Data Mining Community Documentation