Semantic Relatedness Measures for Identifying Relationships in Product Development Processes

Abstract

The Semantic Web, especially in relation to ontologies, provides a structured, formal framework for knowledge interoperability. This trait has been exploited by both the biomedical community in development of the Human Gene Ontology [1] and also by geographers in development of geospatial ontologies [2]. Using semantic relatedness techniques, researchers from both communities have been able to develop and integrate comprehensive knowledge bases. Beyond knowledge integration, semantic relatedness techniques have also been able to provide each community with a unique insight into relationships between concepts in their respective domains. In the engineering community, semantic relatedness techniques promise to provide similar insight into product development processes. This paper explores the application of semantic relatedness techniques to ontologies as a means towards improved knowledge management in product development processes. Several different semantic relatedness techniques are reviewed, including a recently developed meronomic technique specific to domain ontologies. Three of these techniques are adopted to create a semantic relatedness measure specifically designed to identify and rank underlying relationships that exist between aspects of the product development process. Four separate case studies are then presented to evaluate the relative accuracy of the developed algorithm and then determine its effectiveness in exposing underlying relationships.