Developing an ontology is a straightforward process covered by well-known methodologies. As any engineering development, the fundamental basis of the process are the requirements, named competency questions. Traditionally, competency questions are retrieved from interactions with domain experts or by investigating datasets. With the raise in popularity and accessibility of knowledge graphs, we aim to change the paradigm of competency question elicitation. Therefore, in this work, we explore the possibility of extracting competency question directly from Wikidata. We describe in detail the pipeline that makes possible the extraction and abstraction of Wikidata triples, the generation of questions based on triple verbalisations, and the transformation to competency questions. Alongside the method that we apply, the work results in a set of key competency questions from 20 domains and their respective templates mapped to SPARQL query templates.

Watch the webcast replay >>