Cheikh Niang

Using XML data integration and ontology reuse to share agricultural data

We present a data integration approach to enable multi-agency partners to share their agricultural data sources. We perform the integration in three phases. First, a structural integration phase, based on the use of XML documents warehouses (called dataweb), allows to create a warehouse for each partner. A second step consists in the integration of these dataweb by associating an ontology-based knowledge to each warehouse. That is done by a semi-automatic building of OWL ontology starting from dataweb and by re-use of the Agricultural Ontology Service. A third mediation phase permits to query in a uniform manner these different semantic dataweb.

Semantic mediator querying

We present the whole querying process of our ontology-based data integration proposal, that we call Semantic Mediator. The global schema (a TBox) is composed of the source schemas (also Tboxes) and a taxonomy, which links the sources to each other. The querying process is based on the global-schemas structure and consists of three steps: global query rewriting, source querying and global answer building. We describe the overall distributed system and the query-rewriting algorithm. Then we present an application of such a semantic mediation, the Personae project, which is for enabling historians to share their prosopographic data from the Middle Ages and the Renaissance.

Automatic building of an appropriate global ontology

Our objective is to automatically build a global ontology from several data sources, annotated with local ontologies and aiming to share their data in a specific application domain. The originality of our proposal lies in the use of a background knowledge, i.e. a reference ontology, as a mediation support for data integration. We represent ontologies using Description Logics and we combine syntactic-matching with logical-reasoning in order to build the shared global TBox from both the TBoxes of sources and that of the reference ontology.

Towards an Ontology-Based Framework for Data Integration: Application to Agriculture and Health Domains in Senegal

Computer-based systems have reached an advanced level in Africa during the past decade. In Senegal, the agriculture and health domain have led the way by initiating number of projects in ontology-based data integration. In this chapter, we are aiming at two main purposes. First, we present the state of the art in data integration in the agriculture and health domains in Senegal. We focus on the two latest projects: SIC-Senegal for agriculture and SIMENS (the National Medical Information System of Senegal) for the health. Second, we describe the methodological evolution of the ontology-based conceptual framework for data integration stemmed from the lessons learned along the two previous experiences. Finally, we will end this chapter by discussing the new challenges in ontology-based approach in supporting information system design in a larger scale such as climate change risk management. We also discuss the linked open data initiative in Senegal through the KOCC (Knowledge, Ontology, Community and Culture) project.

Querying a Semi-automated Data Integration System

A data integration system enables users to query a unified view of data sources through a global schema. We consider a semi-automatically built data integration system in the semantic web context, where data sources are annotated with ontologies. The global schema is also an ontology, expressed in DL-\(Lite_{\cal A}\). After the semi-automated building of the global schema in a previous work, we focus here on the second part of this system, dedicated to the query answering process. We show how it can rely on either GAV or LAV mappings, that are automatically computed. We present algorithms for both cases, and we discuss the properties of this semi-automatically built data integration system.