Ontology matching is one of the most urgent and important problems on the Semantic Web. In the recent years it became apparent that using existing ontologies to mediate the matching process can have tremendous benefit as compared to the traditional matching methods.
This presentation provides: overview of a framework to perform ontology matching using other ontologies as background knowledge and an insight in matching experiments conducted with existing ontologies. Two ontologies were matched: NALT and Agrovoc, and other six ontologies taken from the Semantic Web were used as background knowledge. The experiments reveal what are the major causes for false matches, and how different characteristics of the background knowledge affect the matching performance.

Watch the webcast replay >>