Data integration provides the ability to manipulate data transparently across multiple distributed data sources. We have studied comprehensively several scenarios where the need for heterogeneous data integration occurs, including centralized integration of XML data sources, hybrid peer-to-peer integration of XML and RDF data sources, pure peer-to-peer XML and RDF data integration and interoperability, personal information management within and across desktops, and geospatial data integration for e-Government.; The key notion of the emerging Semantic Web is that of an ontology , which is a formal and explicit specification of a shared conceptualization. The use of ontologies can benefit data integration tasks in a variety of ways, including metadata representation, global conceptualization, support for high-level queries, declarative mediation, and mapping support. As the main contribution of this thesis, we focus on the role of ontologies in data integration and propose a series of ontology-based approaches to resolve the heterogeneities, including syntactic heterogeneity, schematic heterogeneity, and semantic heterogeneity, so as to achieve data inter operability. In this thesis, we report our achievements on ontology-based heterogeneous data integration, and discuss the fundamental issues, including metadata representation, mapping process, and query processing, in our approaches to different applications of data integration. |