Ontology-based Foundations for Data Integration
Virginija Uzdanaviciute, Rimantas Butleris · 2011
Abstract—The integration of data is one of the most complicated tasks which need to be addressed by IT researchers. Despite the critical importance, the current approaches to semantic interoperability of heterogeneous databases have not been sufficiently effective. We apply an idea of ontology as the foundation for data integration. In the paper the most common data integration methods and their essential features are discussed; the major problems of integration tasks are highlighted. Requirements for implementing data integration tasks are defined for the model, the architecture, the content, and the representation. In order to specify data source semantics, a meta-model is created, which is used to describe the concepts of source and relations. The following architectural ontology-based models are analyzed by selected criteria: a global, a multiple and a hybrid. The model of data integration process is then built upon the hybrid architecture model. Keywords-system interoperability; data integration methods; ontology-based data model; integration approaches; semantic integration I.