An Ontology-Based Framework for Heterogeneous Data Sources Integration

Vânia M. P. Vidal, João C. Pinheiro, Eveline R. Sacramento, José Antonio Fernandes de Macêdo, Bernadette F. Lóscio
2010 Revista de Informática Teórica e Aplicada  
The Proposed Framework Ontologies have been extensively used to model domain-specific knowledge. The main reason for this success is due to their capability to be at the "semantic" level, away from data structures and implementation strategies. In addition, ontology formalisms have allowed certain kinds of reasoning to be automated within a reasonable time complexity. Due to ontology data independence and automated reasoning, ontologies are well suited for integrating heterogeneous databases,
more » ... abling interoperability among disparate systems, and specifying interfaces to independent, knowledge-based services. Recent research has used ontologies for specifying the mediated schema in the context of data integration [2, 3] . An important challenge in ontology-based data integration systems is the problem of rewriting a query specified in terms of the domain ontology into queries that can be answered by the individual data sources. Reasoning is used to determine whether existing ontology concepts (describing the local data sources) are a match for the user's query and to query rewrite. Description Logics can be used to describe relationships between data sources [1] and to provide more flexible mechanisms for semantic query rewriting needed in such systems [2], Despite those approaches, little work has been done focusing on the optimization of querying processing in a ontology-based data integration system. In this work, we propose an ontology-based framework for integration of heterogeneous data sources, comprising the benefits of ontologies to make the semantic integration; and of ontological reasoning to discard sub-queries that are not consistent and to infer additional relations between concepts. Figure 1 describes the main components of the proposed architecture. The mediated schema is represented by a domain ontology (DO), which provides a conceptual representation of the application domain (a global shared vocabulary).
doi:10.22456/2175-2745.12570 fatcat:uuhj4gketnex5ntvcuz2eqtep4