Features of DL-Learner:
- implements different algorithms:
- two refinement operator based algorithms
- a genetic programming algorithm
- a hybrid algorithm using genetic refinement operators
- (random learning, brute force learning)
- supports different kinds of learning problems:
- learning concept definitions and inclusion axioms
- learning from positive and negative examples as well as only from positive examples
- supports different input formats:
- OWL files
- N-triple files
- internal representation in config files
- SPARQL endpoints
- different reasoner adapters:
- DIG interface: allows all major reasoners
- OWL API interface (alpha): FaCT++, Pellet
- different user interfaces:
- command line
- web service
- Java Swing based GUI
- easily extensible through a component model
- 4 types of components: knowledge sources, reasoners, learning problems, learning algorithms
- to implement a new component of one of the above types you only have to extend the correct class in org.dllearner.core and add the name of your file to the components.ini file
- allows a wide range of configuration options