Helping Users Bootstrap Ontologies: An Empirical Investigation

An ontology is a machine processable artifact that captures knowledge about some domain of interest. Ontologies are used in various domains including healthcare, science, and commerce. In this paper we examine the ontology bootstrapping problem. Specifically, we look at an approach that uses both competency questions and knowledge source reuse via recommendations to address the "cold start problem" that is, the task of creating an ontology from scratch. We describe this approach, an implementation of it, and we present an evaluation in the form of a controlled user study. We find that the approach leads users into creating significantly more detailed initial ontologies that have a greater domain coverage than ontologies produced without this support. Furthermore, in spite of a more involved workflow, the usability and user satisfaction of the bootstrapping approach is as good as a state-of-the-art ontology editor with no additional support.