OntoCompare: Assertion Based Comparison of Ontologies on the Semantic Web

With the advent of the Semantic Web, several different ontologies are being used for the addition of semantic annotations to web resources. To facilitate knowledge discovery on the Semantic Web, it has therefore become essential to develop a framework to compare ontologies with minimal human input. Existing Ontology comparison methodologies require a large degree of human intervention and are often inaccurate because they base their comparison solely on the ontological structure ignoring the rich relational semantics present within them. This thesis proposes an assertion based model for comparing ontologies called OntoCompare that reduces human intervention to a minimum and bases the comparison on a holistic view of the ontology that includes the structure as well as the semantic relations. Experimental results on diverse data sets confirm that the OntoCompare scores parallel cognitive comparison performed by human experts thus approximating the human expert's holistic view of the domain.