Advances in modular ontology engineering: methodology and infrastructure

Abstract

Modular ontology engineering is a methodology for producing highly reusable knowledge graph schema. Over the course of this dissertation, we outline a number of contributions that have improved the process to what we see today. These contributions fall within four categories: conveying meaning through schema diagrams, the composition of a modular ontology, the modular ontology engineering methodology, and modular graphical modeling.

First, we created an improved method and tool for generating schema diagrams similar to those manually generated by humans and show that most of OWL, as it is used in real world ontologies, are expressible in this format.

Next, we examined and improved the ontology design pattern development process. This was accomplished through the development of both patterns and modules, extensions to the ontology design pattern representation language, and a tool that significantly improves the usability of these annotations. This work culminated in MODL: a modular ontology design library, which is a distributable set of curated, well-documented ODPs, both novel and drawn from the ontology design pattern portal.

These advances were combined, and building upon the state of the art, to create the Comprehensive Modular Ontology Design IDE (CoModIDE), which is a plugin for the industry-standard ontology editor, Protégé.

Finally, as a culmination of the tool and the methodology, we evaluated CoModIDE, where it was shown to significantly improve outcomes for experienced and new ontology developers when developing modular ontologies.

Altogether, these research topics, resulted in a methodology, that when executed, produced actually reusable, extendable, and adaptable ontologies.