Extraction process specification for materialized ontology views

Abstract

The success of the semantic web relies heavily on ontologies. However, using ontologies for this specific area poses a number of new problems. One of these problems, extracting a high quality ontology from a given base ontology, is currently receiving increasing attention. Areas such as versioning, distribution and maintenance of ontologies often involve this problem. Here, a formalism is presented that enables grouping ontology extraction requirements into different categories, called optimization schemes. These optimization schemes provide a way to introduce quality in the extraction process. An overview of the formalism is discussed, as well as a demonstration of several example optimization schemes. Each of these optimization schemes meets a certain requirement, and consists of rules and algorithms. Examples of how the formalism is deployed to reach a high-quality result, called a materialized ontology view, are covered. The presented methodology provides a foundation for further developments, and shows the possibility of obtaining usable ontologies in a highly automated way.