Software Implementation for Taxonomy Browsing and Ontology Evaluation for the case of Wikidata
Software Implementation for Taxonomy Browsing and Ontology Evaluation for the case of Wikidata
Master's thesis by Serghei Stratan
- Supervisor Markus Krötzsch
- Wissensbasierte Systeme
- 8 Juni 2015 – 7 März 2016
- Download
knowledge represented by relevant terms and formal relations between them, so that machines
can read it automatically. Ontologies on the SemanticWeb come from a vast variety of different
sources, spanning institutions, and persons aiming for different goals and quality criteria.
Measuring some of the aspects of ontologies, enables us to answer the question: “How to assess
the quality of an ontology on the Web?”. The ontology evaluation task is essential for a wide
adoption of ontologies, in the Semantic Web and in other semantically enabled technologies.
The Wikidata project, being a part of a big Wikimedia family, is widely known for its data that is
provided freely to the public. The data is structured and stored in a web ontology. This ontology
has turned immediately into a resource of big value, with a large range of potential applications
across all areas of science, technology, and culture. Unfortunately, at the moment there is no
possibility to assess the quality of the Wikidata ontology.
We have chosen Wikidata as a basic scenario for implementation of an automated system which
would allow to browse the class hierarchy contained in the Wikidata ontology. The class hierarchy
is created by the class entities and two properties (‘subclass of’ and ‘instance of’) which
denote relations between classes. Besides the Wikidata taxonomy browser, additionally we implemented
a component for ontology evaluation, which shows how the quality of the Wikidata
ontology is assessed for several evaluation criteria.