The software is a library as well as a GTK GUI front-end for machine
learning projects. Features:
- Based on intuitionistic fuzzy sets and the possibility theory;
- Features are fuzzy;
- Fuzzy classes, which may intersect and can be treated as features;
- Numeric, enumeration features and ones based on linguistic variables;
- Derived and evaluated features;
- Classifiers as features for building hierarchical systems;
- User-defined features;
- An automatic classification refinement in case of dependent features;
- Incremental learning;
- Object-oriented software design;
- Features, training sets and classifiers are extensible objects;
- Automatic garbage collection;
- Generic data base support (through ODBC);
- Text I/O and HTML routines for features, training sets and classifiers;
- GTK+ widgets for features, training sets and classifiers;
- Examples of use.
This release is packaged for Windows (msi), Fedora (yum) and Debian (apt).
The software is public domain (licensed under GM GPL).
Dmitry A. Kazakov