javaml

Java Machine Learning Library

This document covers the very basic documentation of the library.

The Java Machine Learning Library is licensed under GNU-GPL.

updates: -- 2016/03/03 moved to github.com -- Still working to split GPL code from the core which will be less restrictive, I appreciate pull requests

More elaborate documentation can be found on the web site http://java-ml.sourceforge.net/

1. Overview ============= Java-ML in a nutshell:

   • A collection of machine learning algorithms
   • Common interface for each type of algorithms
   • Library aimed at software engineers and programmers, so no GUI, but clear interfaces
   • Reference implementations for algorithms described in the scientific literature.
   • Well documented source code.
   • Plenty of code samples and tutorials.

2. How to get started ===================== When you are reading this, you most probably already downloaded the library. To use it, include the javaml-.jar in your classpath, as well as the jars that are available in lib/.

How to get started, code samples, tutorials on various tasks can be found at http://java-ml.sourceforge.net

3. Requirements =============== Java 6

4. Dependencies =============== Required libraries:

• Apache Commons Math: used in some algorithms, version 1.2 is included Apache Commons Math is distributed under Apache License 2.0 http://commons.apache.org/math/

• Abeel Java Toolkit: used in some classes, version 2.9 is included AJT is distributed under GNU LGPL 2 or later http://sourceforge.net/projects/ajt/

• Jama: used in some algorithms, version 1.0.2 is included Jama is distributed as public domain software http://math.nist.gov/javanumerics/jama/

Optional libraries:

• Weka: if you like to use algorithms from Weka. Weka 3.6.0 is included in the distribution Weka is distributed under GNU GPL 2 or later http://www.cs.waikato.ac.nz/ml/weka/

• libsvm: if you like to use the libsvm algoriths. libSVM 2.89 is included in this distribution libSVM is distributed under the modified BSD license http://www.csie.ntu.edu.tw/~cjlin/libsvm/

• JUnit: if you want to run the unit tests. As the unit tests are only available from the SVN we did not include a version of JUnit with the distribution. The tests have been written for the JUnit 4 platform and may not function for JUnit 3 or earlier

5. Contact ========== You can contact us by e-mail. [email protected]