jBNC is a Java toolkit for training, testing, and applying Bayesian Network Classifiers. Implemented classifiers have been shown to perform well in a variety of artificial intelligence, machine learning, and data mining applications.
|Tags||education Scientific/Engineering Artificial Intelligence Software Development Libraries Java Libraries|
|Operating Systems||OS Independent|
Release Notes: Binaries of version 1.2 were only compatible with Java 1.5. This version corrects build scripts and features binaries compatible with Java 1.3 and above.
Release Notes: Binaries generated in version 1.2.1 were only compatible with Java 1.5. This version corrects build scripts to generate binaries compatible with Java 1.3 and above.
Release Notes: A serialization interface was implemented to support WEKA 3.4.2, and corresponding additions were made to the test suite. KnowledgeFlow is now supported. The test classes and data were moved to the "test" subdirectory, and a postfix naming convention is now used. The 'weka.bnc' package was moved to 'weka.classifier.bnc' to support the KnowledgeFlow interface, and the GenericObjectEditor.props configuration file was updated to support WEKA 3.4.2.
Release Notes: Serialization was added to some classes to support the use of jBNC in WEKA 3.4.2. This also required some changes to JavaBayes. The Ant build script was modified to first automatically build the updated JavaBayes. Test cases were moved to a separate directory, 'test/src'. The naming convention was changed to postfix.
Release Notes: jBNC-WEKA provides WEKA bindings for jBNC. WEKA is the Waikato Environment for Knowledge Analysis (http://www.cs.waikato.ac.nz/~ml). jBNC-WEKA allows jBNC classifiers and utilities to be run from within WEKA, particularly from within its graphical user interface, called Explorer. Version 1.1 adds support for automatic discretization of continuous attributes. If needed, a custom discretizer is created as part of classifier creation. Discretizers use the jBNC implementation of the entropy discretization algorithm by Fayyad and Irani.