A Method for Building Personalized Ontology Summaries

Abstract

In the context of ontology engineering, the ontology understanding is the basis for its further developmentand reuse. One intuitive e?ective approach to support ontology understanding is the process of ontology summarizationwhich highlights the most important concepts of an ontology. Ontology summarization identi?es an excerpt from anontology that contains the most relevant concepts and produces an abridged ontology. In this article, we present amethod for summarizing ontologies that represent a data source schema or describe a knowledge domain. We proposean algorithm to produce a personalized ontology summary based on user-de?ned parameters, e.g. summary size. Therelevance of a concept is determined through user indication or centrality measures, commonly used to determinethe relative importance of a vertex within a graph. The algorithm searches for the best ontology summary, i.e., theone containing the most relevant ontology concepts respecting all the original relationships between concepts and theparameters set by the user. Experiments were done comparing our generated ontology summaries against goldenstandard summaries as well as summaries produced by methods available in related work. We achieved in average more than 62.5% of similarity with golden standard summaries.