The IGN-E case: Integrating through a hidden ontology

Resumen

National Geographic Institute of Spain (IGN-E) wanted to integrate its main information sources for building a common vocabulary reference and thus to manage the huge amount of information it held. The main problem of this integration is the great heterogeneity of data sources. The Ontology Engineering Group (OEG) is working with IGN-E to attain this objective in two phases: first, by creating automatically an ontology using the semantics of catalogues sections, and second, by discovering mappings automatically that can relate ontology concepts to database instances. So, these mappings are the instruments to break the syntactic, semantic and granularity heterogeneity gap. We have developed software for building a first ontology version and for discovering automatically mappings using techniques that take into account all types of heterogeneity. The ontology contains a set of extra-attributes which are identified in the building process. The ontology, called PhenomenOntology, will be reviewed by domain experts of IGN-E. The automatic mapping discovery will be also used for discovering new knowledge that will be added to the ontology. For increasing the usability and giving independence to different parts, the processes of each phase will be designed automatically and as upgradeable as possible.