Schema Matching Using Duplicates
ICDE '05 Proceedings of the 21st International Conference on Data Engineering
Ontology Matching
| Hi-index | 0.00 |
Ontologies are a useful model to express semantics in a machine-readable way. A matching of heterogeneous ontologies is often required for many different applications like query answering or ontology integration. Many systems coping with the matching problem have been developed in the past, most of them using meta information like concept names as a basis for their calculations. This approach works well as long as the pieces of meta information are similar. In case of very differently structured ontologies or if a lot of possible synonyms, homonyms or meaningless meta information are used, the recognition of mappings gets difficult. In these cases instance-based matching methods are a useful extension to find additional correct mappings resulting in an improved matching quality, because instances provide a lot of information about a concept. This paper presents a novel instance-based matching algorithm which calculates different features using instances. These features characterize the concepts and are compared using different similarity functions. Finally, the similarity values are used to determine 1:1 mapping proposals.