QODI: Query as Context in Automatic Data Integration

QODI is an ontology-based data integration system (OBDI) comprising both ontology mapping and query reformulation methods. QODI is distinguished in that the ontology mapping algorithm dynamically determines a partial mapping specific to the reformulation of each query. The mapping algorithm decomposes a query into a set of paths, and compares the sets of paths with a similar decomposition of a source ontology. The query provides application context not available in traditional ontology matching; thereby the system is able to disambiguate mappings for different queries. The path-based solution also simplifies query reformulation, implicitly resolving the problem of missing entity mappings. Using three real world test sets, QODI achieves favorable results compared to AgreementMaker, a leading ontology matching system, and an ontology-based implementation of the methods detailed for Clio, the state-of-the-art relational data integration and exchange system.