XavierLinNow / jahmm

An implementation of Hidden Markov Models in Java

JAHMM

Jahmm is a Java library implementing the various, well-known algorithms related to Hidden Makov Models (HMMs for short).

The source code of the library is available; it is licensed under GPL (see the resource/COPYING file).

This library is short and simple. It's been written for clarity. It is particularly well suited for research and academic use.

The website associated to this library is: [http://jahmm.googlecode.com/] Most information related to this software can be found there.

This repository is a fork of the original jahmm library that can be found here: [http://jahmm.googlecode.com/]

Compiling

To compile the library, you simply need to compile all the files held in the jahmm/src directory. Thus, simply calling javac with all the .java files held in the jahmm/src directory as arguments compiles everything.

Jahmm requires Java 1.5.0.

Running

To use it, simply launch:

javac -classpath /path/to/jahmm-<version>.jar Myprogram.java

to compile your program, and:

java -cp /path/to/jahmm-<version>.jar myMainClass

(e.g. java -cp /home/smith/java_class/jahmm-0.6.2.jar test/Testing)

...to run it. You can also put the .jar file in your classpath.

Testing

Regression (JUnit) tests have also been written ; see the jahm/test directory.

Files

• pom.xml: the 'maven' project file.
• build.xml: the 'ant' build file.
• src/: all the .java files. src/.../distributions: Pseudo random distributions. src/.../jahmm: The jahmm library itself. This directory holds one directory per java package; see the jahmm website for more information about each of them.
• test/: Regression tests.
• examples: various example files
• README.md: this file.
• CHANGES: changelog.
• ORIGINAL-LICENSE: license file.
• ORIGINAL-THANKS: contributors.

Jutils

The program uses a java library called jutils that can be found here: https://github.com/KommuSoft/jutil

Contact

Jahmm's original author is Jean-Marc Francois.
Feel free to send comments and questions related to this library at:

• http://code.google.com/p/jahmm/issues/list (if you have an issue with the library)
• http://groups.google.com/group/jahmm-discuss or [e-mail] (for questions/comments)

The author of this repository is Willem Van Onsem [e-mail] this version aims to improve speed and enables the use of more sophisticated hidden markov models like the Input-Output Hidden Markov Model (IOHMM). Furthermore decision trees are implemented in the jadetree package.