Automatic semantic enrichment of data services

Abstract

In recent years, many approaches and tools have emerged to assist in the semantic enrichment of Web services. Many researchers have been directing efforts in providing semantic annotations and enriching service descriptions. However, the provisioning of semantically enriched representations through data services has been little considered in the literature, despite of its significant impact on the effectiveness of data integration solutions. In this paper we present a novel method aimed at enriching data provided by services, enabling them to dynamically provide semantic representations. We introduce an architecture, called OntoGenesis, capable of i) constructing and evolving domain ontologies for data services and ii) aligning the properties of such ontologies with external data sources, so as to enhance the reuse of well-known concepts in order to facilitate further integrations. Experiments using real-world bilingual datasets show the applicability of our proposal and the obtained results reveal that alignments achieve satisfactory levels of precision and recall.