Protégé DL-Learner Plugin

The Protégé 4 DL-Learner plugin allows to learn equivalence and super class axioms based on the instance data in the ontologies loaded in Protégé. It integrates seamlessly in Protégé.

You can either download the plugin or install directly from Protégé 4 itself: Go to “File” => “Preferences” => “Plugins”, click on “Check for Downloads” and select the DL-Learner plugin. Protege 4 build 111 is required.

The following screen cast shows how to activate the plugin in Protégé and use it:

View Screencast

The plugin was created by Christian Kötteritzsch and Jens Lehmann. If you have any questions or comments, please contact me.

Features

• learns definition and super class axioms based on existing instance data in an ontology
• seamless integration through use of Protégé 4 plugin mechanism
• based on DL-Learner Machine Learning framework to generate suggestions
• use of efficient and sophisticated machine learning algorithms
• can handle cases where an ontology is already damaged, i.e. adding the desired axiom renders the ontology inconsistent – in this case it displays the instances causing inconsistency

ChangeLog

• Version 0.5.2:
  • fixed bugs #2790483 and #2791135, corrected manifest file, ensured Java 5 compatibility
• Version 0.5.1:
  • corrected display of suggested class expressions; message when learning algorithm is still running
• Version 0.5:
  • better integration in Protege (thanks to Nick Drummond for his support)
  • switch from general machine learning algorithm to special purpose algorithm for definitions and super class axioms
  • better performance through stochastic coverage testing procedures in machine learning algorithm
  • visualisation of learned results in Protege to simplify the decision whether to add the axiom for the knowledge engineer
• Version 0.1: initial release

Screenshot:

See also the DL-Learner page in the Protege wiki.