Ontology Matching Techniques to Support Applications in Open Environments

In the context of the OpenKnowledge project we have developed two ontology matching techniques that support tools operating in open environments. Such tools typically re-use multiple, heterogeneous ontologies identified at runtime. The first algorithm, PowerMap is the core mapping component of the PowerAqua question answering system and enables a flexible mapping of a user query to dynamically selected ontologies from the Semantic Web. As such, PowerMap is a key component for bridging the semantic gap between the user’s terminology and the various conceptualisations used in online ontologies.

Our second algorithm, Scarlet, discovers semantic relations between the concepts of two ontologies by using the entire Semantic Web as its background knowledge. Scarlet derives relations by combining multiple facts that are dynamically identified and extracted from online ontologies. Therefore, this algorithm provides an added benefit in scenarios where appropriate background ontologies cannot be identified manually. For example, besides matching, we have used Scarlet for supporting tasks such as the enrichment of folksonomy tagspaces and ontology evolution.

More information is available in Deliverables 3.2, 8.4 and 8.5.

Related Publications:

Egglue Powered