Conceptual Classification and Domain Ontology for Semantic Integration of Data

Abstract

The multiplicity of data sources, often heterogeneous, like the ones used for the web renews interest in data integration. In this context, we present a solution that searches multiple data sources and extracts relevant information as a response to a client query. Our solution can be either seen as an internal component of a mediation system or as a stand-alone information retrieval system.

The data sources we have considered are all related to the same domain; which is “Tourism”. We annotated them according to this domain ontology. The search method we adopted used a clustering technique based on concept lattices. This algorithm allows formal classification of data sources and the results it returns are ordered by degree of pertinence.

To validate our solution, we implemented it as a Java application for a case study with about thirty data sources and twenty terms issued from the ontology.